S PN=5577239 OR PN=5950192 OR PN=4642762
2 PN=5577239
1 PN=5950192
1 PN=4642762
S2 3 PN=5577239 OR PN=5950192 OR PN=4642762
?
TYPE 2/2/ALL
2/2/1 (Item 1 from file: 653)
DIALOG(R)File 653:US Patents Fulltext
(c) format only 2001 The Dialog Corp. All rts. reserv.
01566805
Utility
STORAGE AND RETRIEVAL OF GENERIC CHEMICAL STRUCTURE REPRESENTATIONS
PATENT NO.: 4,642,762
ISSUED: February 10, 1987 (19870210)
INVENTOR(s): Fisanick, William, Columbus, OH (Ohio), US (United States of
America)
ASSIGNEE(s): American Chemical Society, (A U.S. Company or Corporation ),
Washington, DC (District of Columbia), US (United States of
America)
EXTRA INFO: Expired, effective February 10, 1999 (19990210), recorded in
O.G. of April 20, 1999 (19990420)
Reinstated, effective June 28, 1999 (19990628), recorded in
O.G. of July 27, 1999 (19990727)
APPL. NO.: 6-614,219
FILED: May 25, 1984 (19840525)
U.S. CLASS: 707-3 cross ref: 707-104
INTL CLASS: [4] G06F 15-40
FIELD OF SEARCH: 364-200MSFILE; 364-900MSFILE
References Cited
U.S. PATENT DOCUMENTS
4,473,890 9/1984 Araki 364-900
PRIMARY EXAMINER: Zache, Raulfe B.
ATTORNEY, AGENT, OR FIRM: Pollick, Philip J.
CLAIMS: 31
EXEMPLARY CLAIM: 1
DRAWING PAGES: 22
DRAWING FIGURES: 22
ART UNIT: 232
FULL TEXT: 1514 lines
2/2/2 (Item 1 from file: 654)
DIALOG(R)File 654:US PAT.FULL.
(c) format only 2001 The Dialog Corp. All rts. reserv.
02999303
Utility
RELATIONAL DATABASE MANGEMENT SYSTEM FOR CHEMICAL STRUCTURE STORAGE,
SEARCHING AND RETRIEVAL
PATENT NO.: 5,950,192
ISSUED: September 07, 1999 (19990907)
INVENTOR(s): Moore, Jeffrey, Timonimun, MD (Maryland), US (United States of
America)
Brazil, Joanne, White Hall, MD (Maryland), US (United States
of America)
Hoover, Jeffrey R., Baltimore, MD (Maryland), US (United
States of America)
ASSIGNEE(s): Oxford Molecular Group, Inc , (A U.S. Company or Corporation),
Towson, MD (Maryland), US (United States of America)
APPL. NO.: 8-883,165
FILED: June 26, 1997 (19970626)
This application is a continuation, of application Ser. No. 08-715,708,
filed Sep. 19, 1996, now abandoned, which is a continuation application of
Ser. No. 08-288,503, filed Aug. 10, 1994, now U.S. Pat. No. 5,577,239, the
entire disclosure of which is incorporated herein by reference.
U.S. CLASS: 707-3 cross ref: 702-27
INTL CLASS: [6] G06F 17-30
FIELD OF SEARCH: 395-496; 395-497; 395-499; 395-600; 395-603; 707-3; 707-2;
707-1; 707-100; 707-102; 707-104; 707-22; 707-27; 707-19;
707-20; 702-22; 702-27; 702-19; 702-20
References Cited
U.S. PATENT DOCUMENTS
4,642,762 2/1987 Fisanick 707-3
4,811,217 3/1989 Tokizane et al. 364-300
4,855,931 8/1989 Saunders 364-499
5,025,388 6/1991 Cramer, III et al. 364-496
5,056,035 10/1991 Fujita 364-497
5,259,137 11/1993 Wilson et al. 364-496
5,367,058 11/1994 Pitner et al. 530-391.9
5,379,234 1/1995 Wilson et al. 364-496
5,386,507 1/1995 Teig et al. 395-161
5,418,944 5/1995 DiPace et al. 395-600
5,463,564 10/1995 Agrafiotis et al. 364-496
5,577,239 11/1996 Moore et al. 395-603
NON-U.S. PATENT DOCUMENTS
090 895 A2 10/1983 EP (European Patent Office)
213 483 A2 3/1987 EP (European Patent Office)
OTHER REFERENCES
Viking Instruments Corp. (Hewlett Packard); SpectraTrak Transportable GS/MS
Systems; (brochure)-No Date.
Chemical Structures, The International Language of Chemistry; Wendy A. War
(Ed.); "Interfacing DARC--Oracle" AJCM (Juus) de Jong (1988).
J. Chem. Inf. Comput. Sci. (1983) , vol. 23, No. 3; pp. 102-108; DARC
Substructure Search System: A New Approach to Chemical Information; Roger
Attias.
J. Chem. Inf. Comput. Sci. (1987), vol. 27, No. 2; pp. 74-82; DARC System:
Notions of Defined and Generic Substructures. Filiation and Coding of FREL
Substructure (SS) Classes; Jacques-Emile Dubois et al.
J. Chem. Inf. Comput. Sci. (1990), vol. 30, No. 2; pp. 191-199,
Substructure Search Systems. 1. Performance Comparison of the MACCS, DARC,
HTSS, and CAS Registry MVSSS, and S4 Substructure Search System; Martin G.
Hicks & Clemens.
J. Chem. Inf. Comput. sci. (1988), vol. 28, No. 4; pp. 221-226; An
Efficient Graph Approach to Matching Chemical Structures, O. Owolabi.
J. Chem. Inf. Comput. Sci. (1990), vol. 30, No. 4; pp. 332-339; Reactions
in the Beilstein Information System: Nonaporic Organic Synthesis; Martin G.
Hicks.
Analytica Chimica Acta, 235 (1990), pp. 87-92; Substructure Search Systems
for Large Chemical Data Bases; Martin G. Hicks et al.
J. Chem. Inf. Comput. Sci. (1991), vol. 31, No. 2; pp. 320-326; The
Beilstein Structure Registry System. 1. General Design; Laszio Domokos.
J. Chem. Inf. Comput. Sci. (1989), vol. 29, No. 4; pp. 255-260; 3DSearch; A
System for Three-Dimensional Substructure Searching; Robert P. Sheridan, et
al.
Substructure Searches of Chemical Structure Files; (Jan. 23, 1973);
Strategic Considerations in the Design of a Screening System for
Substructure Searches of Chemical Structure Files; George W. Adamson, et
al.
Chemical Structure Searching; (Jan. 21, 1975); An Efficient Design for
Chemical Structure Searching. I. The Screens; Alfred Feldman et al.
J. Chem. Inf. Comput. Sci. (1982), vol. No. 4; The Third BASIC Fragment
Search Dictionary; W. Graf, H. K. Kaindl, et al.
J. Chem. Inf. Comput. Sci. (1983), vol. 23, No. 3; The CAS Online Search
System. 1. General System Design and Selection, Generation, and Use of
Search Screens; P. G. Dittmar, et al.
Computer Chemical, ((1991), vol. 15, No. 2, pp. 103-107; A Central Atom
Based Algorithm and Computer Program for Substructure Search; Alf Dengler
and Ivar Ugi.
J. Chem. Inf. Comput. Sci. (1993), vol. 33, No. 4; pp. 545-547; Sturcture
Searching in Chemical Databases by Direct Lookup Methods; Baradley D.
Christie et al.
PRIMARY EXAMINER: Von Buhr, Maria N.
ATTORNEY, AGENT, OR FIRM: Dickstein Shapiro Morin & Oshinsky
CLAIMS: 14
EXEMPLARY CLAIM: 1
DRAWING PAGES: 7
DRAWING FIGURES: 12
ART UNIT: 277
FULL TEXT: 798 lines
2/2/3 (Item 2 from file: 654)
DIALOG(R)File 654:US PAT.FULL.
(c) format only 2001 The Dialog Corp. All rts. reserv.
02592280
Utility
CHEMICAL STRUCTURE STORAGE, SEARCHING AND RETRIEVAL SYSTEM
PATENT NO.: 5,577,239
ISSUED: November 19, 1996 (19961119)
INVENTOR(s): Moore, Jeffrey, 12 Breezy Tree Ct., Timonimun, MD (Maryland),
US (United States of America), 21093
Brazil, Joanne, 4500 Jolly Acres Rd., White Hall, MD
(Maryland), US (United States of America), 21161
Hoover, Jeffrey R., 8639 Willow Oak Rd., Baltimore, MD
(Maryland), US (United States of America), 21234
[Assignee Code(s): 68000]
EXTRA INFO: Assignment transaction [Reassigned], recorded October 12,
1994 (19941012)
Assignment transaction [Reassigned], recorded January 21,
1997 (19970121)
APPL. NO.: 8-288,503
FILED: August 10, 1994 (19940810)
U.S. CLASS: 707-3 cross ref: 702-27
INTL CLASS: [6] G06F 17-30
FIELD OF SEARCH: 364-DIG.1; 364-DIG.2; 364-496; 364-497; 364-499; 395-600
References Cited
U.S. PATENT DOCUMENTS
4,642,762 2/1987 Fisanick 364-300
4,811,217 3/1989 Tokizane et al. 364-300
4,855,931 8/1989 Saunders 364-499
5,025,388 6/1991 Cramer, III et al. 364-496
5,056,035 10/1991 Fujita 364-497
5,249,137 2/1993 Wilson et al. 364-496
5,367,058 11/1994 Pitner et al. 530-391.9
5,379,234 1/1995 Wilson et al. 364-496
5,386,507 1/1995 Teig et al. 395-161
5,418,944 5/1995 DiPace et al. 395-600
5,463,564 10/1995 Agrafiotis et al. 364-496
OTHER REFERENCES
Viking Instruments Corp. (Hewlett Packard); Spectra Trak Transportable
GC/MS System; (brochure), No date.
Chemical Structure, The International Language of Chemistry; Wendy A. War
(Ed.); "Interfacing DARC-Oracle" AJCM (Juus) de Jong (1988).
J. Chem. Inf. Comput. Sci. (1983), vol. 23, No. 3 pp. 102-108; DARC
Substructure Search System; A New Approach to Chemical Information; Roger
Attias.
J. Chem. Inf. Comput. Sci. (1987), vol. 27, No. 2; pp. 74-82; DARC System;
Notions of Defined and Generic Substructures. Filiation and Coding of FREL
Substructure (SS) Classes; Jacques-Emile Dubois et al.
J. Chem. Inf. Comput. Sci. (1990), vol. 30, No. 2; pp. 191-199,
Substructure Search Systems, 1, Performance Comparison of the MACCS, DARC,
HTSS, CAS Registry MVSSS, and S4 Substructure Search Systems; Martin G.
Hicks.
J. Chem. Inf. Comput. Sci. (1988), vol. 28, No. 4; pp. 221-226; An
Efficient Graph Approach to Matching Chemical Structures, O. Owolabi.
J. Chem. Inf. Comput. Sci. (1990), vol. 30, No. 4; pp. 332-339; Reactions
in the Bellstein Information System: Nonaporic Organic Synthesis; Martin G.
Hicks.
Analytica Chimica Acta, 235 (1990), pp. 87-92; Substructure Search Systems
for Large Chemical Data Bases; Martin G. Hicks et al.
J. Chem. Inf. Comput. Sci. (1991), vol. 31, No. 2; pp. 320-326; The
Bellstein Structure Registry System, 1, General Design; Laszio Domokos.
J. Chem. Inf. Comput. Sci. (1989), vol. 29, No. 4; pp. 255-260; 3DSearch; A
System for Three-Dimensional Substructure Searching; Robert P. Sheridan, et
al.
Substructure Searches of Chemical Structure Files; (Jan. 23, 1973);
Strategic Considerations in the Design of a Screening System for
Substructure Searches of Chemical Structure Files; George W. Adamson, et
al.
Chemical Structure Searching; (Jan. 21, 1975); An Efficient Design for
Chemical Structure Searching, I, The Screens; Alfred Feldman et al.
J. Chem. Inf. Comput. Sci. (1982), vol. 22, No. 4; The Third BASIC Fragment
Search Dictionary; W. Graf, H. K. Kaindl, et al.
J. Chem. Inf. Comput. Sci. (1983), vol. 23, No. 3; The CAS ONLINE Search
System, 1, General System Design and Selection, Generation, and Use of
Search Screens; P. G. Dittmar, et al.
Computer Chemical, (1991), vol. 15, No. 2; pp. 103-107; A Central Atom
Based Algorithm and Computer Program for Substructure Search; Alf Dengler
and Ivar Ugi.
J. Chem. Inf. Comput. Sci. (1993), vol. 33, No. 4; pp. 545-547; Structure
Searching in Chemical Databases by Direct Lookup Methods; Bradley D.
Christie et al.
J. Chem. Inf. Comput. Sci. (1993); vol. 33, No. 4; pp. 539-541;
Substructure Searching on Very Large Files by Using Multiple Storage
Techniques; Alexander Bartmann et al.
PRIMARY EXAMINER: Black, Thomas G.
ASST. EXAMINER: Von Buhr, Maria N.
ATTORNEY, AGENT, OR FIRM: Dickstein Shapiro Morin & Oshinsky LLP
CLAIMS: 12
EXEMPLARY CLAIM: 1
DRAWING PAGES: 7
DRAWING FIGURES: 12
ART UNIT: 237
FULL TEXT: 791 lines
?
TYPE 2/2,EM,SU/ALL
2/2,EM,SU/1 (Item 1 from file: 653)
DIALOG(R)File 653:US Patents Fulltext
(c) format only 2001 The Dialog Corp. All rts. reserv.
01566805
Utility
STORAGE AND RETRIEVAL OF GENERIC CHEMICAL STRUCTURE REPRESENTATIONS
PATENT NO.: 4,642,762
ISSUED: February 10, 1987 (19870210)
INVENTOR(s): Fisanick, William, Columbus, OH (Ohio), US (United States of
America)
ASSIGNEE(s): American Chemical Society, (A U.S. Company or Corporation ),
Washington, DC (District of Columbia), US (United States of
America)
EXTRA INFO: Expired, effective February 10, 1999 (19990210), recorded in
O.G. of April 20, 1999 (19990420)
Reinstated, effective June 28, 1999 (19990628), recorded in
O.G. of July 27, 1999 (19990727)
APPL. NO.: 6-614,219
FILED: May 25, 1984 (19840525)
U.S. CLASS: 707-3 cross ref: 707-104
INTL CLASS: [4] G06F 15-40
FIELD OF SEARCH: 364-200MSFILE; 364-900MSFILE
References Cited
U.S. PATENT DOCUMENTS
4,473,890 9/1984 Araki 364-900
PRIMARY EXAMINER: Zache, Raulfe B.
ATTORNEY, AGENT, OR FIRM: Pollick, Philip J.
CLAIMS: 31
EXEMPLARY CLAIM: 1
DRAWING PAGES: 22
DRAWING FIGURES: 22
ART UNIT: 232
FULL TEXT: 1514 lines
FIELD
This invention relates to a method for storing and retrieving generic
chemical structure representations (Markush formulations) and information
associated with them. It is directed especially to development of
specific(real)-atom and generic-group representations of the Markush
formulation that are used in atom-by-atom and group-by-group comparison of
query and file representations and the use of screening techniques to
eliminate a high percentage of irrelevant file representations prior to
group-by-group and atom-by-atom comparison of generic-group and
specific-atom representations.
BACKGROUND
The ability to effectively retrieve information on generic chemical
structures, i.e., so-called Markush structures, has been a problem of
varying magnitude and complexity since the inception of the use of the
Markush claim by the Patent Office in the 1920's. Many manual and
mechanized information retrieval systems have been developed to meet the
challenge of this problem but the known techniques for such retrieval are
imprecise and often place a premium on the knowledge, intuition, and
cognitive skills of the searcher.
The basic system for dealing with Markush structures is a manual system
in which individual documents containing the Markush structure are
classified according to a highly refined classification system and
physically grouped according to the classification scheme into a search
file. In making a search, the searcher proceeds by classifying the document
(query) in hand and then goes to the appropriately classified physical
group of documents in the search file and manually searches those documents
for relevant retrievals. Such a system places a high premium on the correct
initial classification of search file documents, correct classification of
the query, physical search-file integrity, and highly-developed cognitive
skills of the searcher. Moreover, because the Markush may represent
thousands or even millions of compounds, it often is impossible to
promulgate copies of the document into all of the search file
classifications represented by the Markush formulation. Weaknesses in any
of the aforementioned areas is likely to produce unsatisfactory search
results. (U.S. Department of Commerce, "Development and Use of Patent
Classification Systems", U.S. Government Printing Office, Washington, D.C.,
1966.)
Another technique used in both manual and mechanized systems for the
handling of Markush structures involves the use of a system of
fragmentation codes that are in effect generic or real-atom "group"
representations of portions of a particular Markush formulation. For
example, that portion of the formulation containing chains of carbon atoms
might be generically encoded as alkyl, or OH group as an alcohol or
hydroxide, and F, Cl, Br, and I as a halide. Real-atom groups, such as
methyl for CH sub 3 13 , ethyl for CH sub 3 CH sub 2 --, and phenyl for C
sub 6 H sub 5 --, are also typically used. (Balent, M. Z.; Emberger, J. M.
"A Unique Chemical Fragmentation System for Indexing Patent Literature" J.
Chem. Inf. Comput. Sci. 1975, 15, 100-104. Kaback, S. M. "Chemical
Structure Searching in Derwent's World Patents Index" J. Chem. Inf. Comput.
Sci. 1980, 20, 1-6. Rossler, S.; Kolb, A. "The GREMAS System, an Integral
Part of the IDC System for Chemical Documentation" J. Chem. Doc. 1970, 10,
128-134. Rowlett, R. J. "Gleaning Patents with Chemical Abstracts" Chemtec.
1979, June, 348-349. Silk, J. A. "Present and Future Prospects for
Structural Searching of the Journal and Patent Literature." J. Chem. Inf.
Comput. Sci. 1979, 19, 195-198.) However, the inter-relationships among
these groups in a Markush formulation are typically not encoded. As a
result, such systems tend to have good recall, i.e., most of the relevant
search file answers are retrieved but, because the inter-relationship among
the groups can not be specified and the reliance on generic terminology,
such systems have a pronounced tendency to lack precision, i.e., many of
the answers retrieved are irrelevant to the query. Precision has been
improved by incorporation of a higher degree of specificity into the
fragmentation codes, but only at a price paid in terms of higher complexity
and difficulty in file encoding and search profile formulation and a
resulting higher potential for error.
Mechanized specific atom-by-atom structure matching of query and file
structural representations is a well-known commercial technique that has
been available since the 1960s and has demonstrated high recall and
precision as a search and retrieval technique. (Wigington, R. L. "Machine
Methods for Accessing Chemical Abstracts Service Information in Proceedings
of the IBM Symposium on Computers and Chemistry"; IBM Data Processing
Division: White Plains, NY, 1969. Eakin, D. R. "The ICI CROSSBOW System,"
in Ash, J. E.; Hyde, E., Eds. Chemical Information Systems, Chichester,
Horwood, 1975. Dubois, J. E. "DARC System in Chemistry", in Computer
Representation and Manipulation of Chemical Information, Wipke, W. T.;
Heller, S.; Feldman, R.; Hyde, E., Eds., Wiley, New York, 1974. Schenk, H.
R.; Wegmuller, F. "Substructure Search by Means of the Chemical Abstracts
Service Chemical Registry II System" J. Chem. Inf. Comput. Sci. 1976, 16,
153-161. Feldman, R. J. "Interactive Graphic Chemical Substructure
Searching" in Computer Representation and Manipulation of Chemical
Information, Wipke, W. T.; Heller, S.; Feldman, R.; Hyde, E., Eds., Wiley,
New York, 1974.) Because atom-by-atom structure matching is a relatively
slow process, screening techniques have been developed to eliminate a high
percentage of irrelevant file representations. Typically screening involves
capturing key features of the file representations such as atom environment
and atom sequences and then matching similar key features of the query
representation to give a set of answers that are then used in atom-by-atom
structure matching. (Dittmar, P. G.; Farmer, N. A.; Fisanick, W.; Haines,
R. C.; Mockus, J. "The CAS ONLINE Search System. 1. General System Design
and Selection, Generation, and Use of Search Screens" J. Chem. Inf. Comput.
Sci. 1983, 23, 93-102. Attias, R. "DARC Substructure Search System: A New
Approach to Chemical Information" J. Chem. Inf. Comput. Sci. 1983, 23,
102-108.) Unfortunately, structure matching techniques tend to be limited
to files containing representations of unique individual compounds and
queries have been limited to specific structural representations that must
exactly match the structural representation of the file compound
(full-structure search) or be embedded within it (substructure search).
Structure matching techniques have been applied to Markush formulations
which represent a relatively small number of specific compounds using
queries that contain only real atoms. (Meyer, E. "Topological Search for
Classes of Compounds in Large Files--even of Markush Formulas--at
Reasonable Machine Cost" in Computer Representation and Manipulation of
Chemical Information, Wipke, W. T.; Heller, S.; Feldman, R.; Hyde, E.,
Eds., Wiley, New York, 1974.) However, in attempting to apply structure
matching techniques to query and file structures represented by Markush
formulations of the type often found in broad patent claims, one is
immediately faced with the problem that a single Markush formulation may
literally represent millions of specific compounds. When one considers that
the file size of the current large commercial structural matching systems
is a little less than seven million specific compounds, an appreciation is
gained for the difficulty in using structure matching techniques to search
effectively Markush structures. Although proposals have been made to apply
structure matching techniques to broad Markush formulations, no viable
system for searching such Markush formulations that gives a high degree of
recall and precision has yet been achieved. (Lynch, M. F.; Bernard, J. M.;
Welford, S. M. "Computer Storage and Retrieval of Generic Chemical
Structures in Patents. 1. Introduction and General Strategy" J. Chem. Inf.
Comput. Sci. 1981, 21, 148-150. Barnard, J. M.; Lynch, M. F.; Welford, S.
M. "Computer Storage and Retrieval of Generic Chemical Structures in
Patents. 2. GENSAL, a Formal Language for the Description of Generic
Chemical Structures" J. Chem. Inf. Comput. Sci. 1981, 21, 151-161. Welford,
S. M.; Lynch, M. F.; Barnard, J. M. "Computer Storage and Retrieval of
Generic Chemical Structures in Patents. 3. Chemical Grammars and their Role
in the Manipulation of Chemical Structures" J. Chem. Inf. Comput. Sci.
1981, 21, 161-168. Barnard, J. M.; Lynch, M. F.; Welford, S. M. "Computer
Storage and Retrieval of Generic Chemical Structures in Patents. 4. An
Extended Connection Table Representation (ECTR) for Generic Structures." J.
Chem. Inf. Comput. Sci. 1982, 22, 160-164. Nakayama, T.; Fujiwara, Y.
"Computer Representation of Generic Chemical Structures by an Extended
Block-Cutpoint Tree" J. Chem. Inf. Comput. Sc 1983, 23, 80-87. Kudo, Y.;
Chihara H. "Chemical Substance Retrieval System for Searching Generic
Representations. 1. A Prototype System for the Gazetted List of Existing
Chemical Substances of Japan" J. Chem. Inf. Comput. Sci. 1983, 23,
109-117.)
SUMMARY
A typical Markush storage and retrieval process according to the resent
invention comprises the steps of forming a file of structural
representations of Markush formulations in which each Markush formulation
is represented by a single specific atom multiple connectivity node (SpMCN)
representation in which the formal valance requirements of requisite atoms
are relaxed to allow for the attachment of all atoms and groups of atoms
depicted in the Markush formulation and, as a result, gives a
representation containing all implicit specific atom structures found in
the Markush formulation. The SpMCN is then converted to an associated
generic group multiple connectivity node (GnMCN) representation through the
use of a generic-group hierarchy. A query Markush formulation is similarly
converted to SpMCN and GnMCN representations. The query GnMCN
representation then is compared on a group-by-group basis with each file
GnMCN in such fashion so that a match is found when at least one implicit
generic structure representation (IGSR) of the query GnMCN is identical
with (overlaps) or is contained in (embedded in) at least one IGSR of the
file GnMCN. The query SpMCN representation then is compared on an
atom-by-atom basis with the file SpMCN representations associated with the
file GnMCN representations (answers) obtained in the previous query
GnMCN/file GnMCN matching step in such fashion so that a match is found
when at least one implicit specific atom structure representation (ISSR) of
the query SpMCN structure is identical with (overlaps) or is contained in
(embedded in) at least one ISSR of the file SpMCN. An indexing system is
used to identify IGSRs and ISSRs for the matching process and to manipulate
large or complex GnMCN and SpMCN representations.
As a further refinement, generic features of the original Markush
formulation are captured by using the generic-group hierarchy as a means of
representing generic features of the Markush formulation in both the SpMCN
and GnMCN. To insure high recall, a roll-back feature is used to allow for
the exchange of generic-group and specific-atom representations in SpMCN
matching so that all real atom file or query structural features implied in
the generic structural features of the file or query SpMCN are matched. In
addition, specific features of the SpMCN and specifically identified parts
of generic features of the original Markush formulation, such as specific
atoms, type of bonding, ring size, etc. are associated with each generic
group of the file GnMCN as group attributes and are matched against group
attributes of the generic groups of the query GnMCN prior to SpMCN
matching.
As a further refinement, screening techniques are applied to both SpMCN
and GnMCN representations in order to eliminate a large number of
irrelevant file representations prior to the more exacting group-by-group
and atom-by-atom comparisons. In order to achieve a high level of recall, a
Boolean strategy is used in the query screen logic expression whereby
special, "diagnostic" generic-group screens are used as alternatives for
sets of specific-atom screens in order to retrieve file answers in
situations where real-atom structures of the SpMCN query structure are
implied in the generic portions of the file SpMCNs that originate from
generic features of the original Markush formulation and for which there
are no real-atom counterparts.
DETAILED DESCRIPTION
A simple Markush formulation is set forth in structure Ia of FIG. 1. This
formulation consists of a fixed structure portion to which is attached the
variable groups R sub 1 and R sub 2. As indicated in the text portion of
the formulation, R sub 1 may be chlorine (Cl) or bromine (Br) and R sub 2
may be ethyl (CH sub 3 CH sub 2) or methyl (CH sub 3). Implicit in the
Markush formulation is the representation of four distinct individual
compound representations, Ia1-Ia4, that are, in effect, all of the possible
individual structures resulting from the combinations of fragments in the
variable groups denoted by R sub 1 and R sub 2.
In representation Ia, it is noted that carbon (C) typically has a
connectivity (valance) of four, i.e., is capable of attaching or connecting
itself to four other entities or to fewer than four other entities via a
multiple bond to one or more of the entities. Specifically in the ring
system of representation Ia, each carbon is bound to a second carbon by a
multiple (double) bond, to a third carbon atom by a single bond, and to a
hydrogen atom (H) or to a variable group node (R sub 1,R sub 2) by a single
bond to give the usual carbon valance of four. As is shown in structures
Ia1-Ia4, it is common practice in the chemical arts often to omit the
hydrogen atoms and to designate the alternate single and double bonds
between carbon atoms in the ring as a circle, the later convention is felt
to represent more realisticly a delocalized bonding situation in which
there are more like one and a half bonds between all carbon atoms. Except
where noted, these common conventions will be followed throughout the
remainder of the specification and drawings.
Structure Ib of FIG. 1 is a multiple connectivity node (MCN) structure.
In it, all of the fragments belonging to the variable groups described in
the text part of the Markush formulation have been attached to their
respective nodes or points of variability (as shown in the structure part
of the Markush formulation Ia) giving rise to nodes of abnormally high
connectivity and hence the multiple connectivity node (MCN) designation.
Since Ib represents all of the specific atoms identified in the Markush
formulation, it is designated as a specific-atom multiple connectivity node
(SpMCN) representation. It should be noted that the four distinct
individual compound representations, Ia1-Ia4, are also implicit in the
SpMCN. These individual implicit representations are referred to as
implicit specific-atom structural representations (ISSRs).
By using common generic technology, it is possible to simplify further
the specific multiple connectivity node structure (SpMCN). For example,
carbon ring structures containing only carbon atoms in the ring are often
given the generic description of carbocycles; linear chains of carbon atoms
are generically termed alkyls; and chlorine and bromine are often called
halides. Using this basic generic terminology, it is possible to transform
the SpMCN representation shown in Ib to the generic multiple connectivity
node representation (GnMCN) shown in Ic. In transforming the SpMCN to a
GnMCN, the bonding level between the generic groups is preserved. In this
particular example, only a single bond exists between the carbocycle and
the variable groups. If, however, a multiple bond exists between the
generic representations, such bonding is indicated in the GnMCN. Implicit
within the GnMCN representation are four implicit generic group structure
representations (IGSRs), Ic1-Ic4, corresponding to the four distinct
compound representations, ISSRs, implicit in the original Markush and the
SpMCN. It is critical to note that IGSRs and ISSRs are used for
illustrative purposes only. This invention does not anticipate the storage
of all ISSRs and IGSRs associated with the respective SpMCN and GnMCN.
Rather the invention is directed at the indivdiual ISSRs and IGSRs as they
are implicitly contained within the SpMCN and GnMCN. The actual
representation and processing uses only the explicit SpMCN and GnMCN
representations. The ISSRs and IGSRs are used only as they are implicitly
found within the SpMCN and GnMCN representations.
FIGS. 2, 2', 3, 3' and 4 illustrate the use of the GnMCN and SpMCN
representations in file searching and retrieval. In FIGS. 2 and 2',
representations IIa-VIa are illustrative file Markush formulations as they
might appear in patent documents, IIb-VIb are SpMCN representations of the
corresponding Markush formulation, and IIc-VIc are GnMCN representations of
the corresponding SpMCN representations. The query Markush formulation VIIa
is also shown as a SpMCN representation (VIIb) and a GnMCN representation
(VIIc). As shown in FIGS. 3 and 3', a file search is initiated by comparing
each query IGSR (VIIc1 and VIIc2) with each file IGSR (11c1-IIc5,
IIIc1-IIIc3, IVc1-IVc3, Vc1-Vc4, and VIc1-VIc4; identical implicit
structures are shown only once). As seen, query IGSR VIIc1 matches with
file IGSRs IIc1-IIc5 and Vc4; VIIc2 matches with Vc2-Vc3 and VIc1-VIc4. At
this point, representations III and IV have been eliminated from the search
and, as shown in FIG. 4, matching now proceeds between the query ISSRs
VIIb1-VIIb2 and the file ISSRs IIb1-IIb6, Vb1-Vb4, and VIb1-VIb4; query
ISSR VIIb1 matches only with file ISSR IIb1 and query ISSR VIIb2 matches
nothing, specifically illustrating that only one ISSR need match one file
ISSR to give an answer. To complete the search, relevant information
associated with the Markush formulation IIa such as, but not limited to, an
abstract, patent number, or patent document is retrieved for the searcher.
FIG. 5 illustrates the two types of matching criteria that a searcher may
use in carrying out a search. Representation VIII is a single compound
representation in which the ISSR is identical with the actual structure.
This structure matches exactly with the file representation X which is also
a single compound representation. The exact matching of all characteristics
of the query representation with those of the file representation is termed
"overlap" or full-structure search. Exact matching may be relaxed such that
the query representation need only be contained within the file
representation. Thus, although query representation VIII does not exactly
match or "overlap" the single file representation XI, it is contained
within representation XI. Such containment of the query representation
within the file representation is termed "embedment" or substructure
search. Systems for both full-structure and substructure search are
commercially available, e.g., CAS ONLINE: The Registry File, Chemical
Abstracts Service, Columbus, Ohio.
Atom-by-atom searching involves the comparison of a query structure with
a file structure using a path-tracing technique. Typically the path-tracing
technique involves selecting a starting atom (node) of the query structure
(usually a noncarbon atom) and comparing it with the first atom of the file
structure. If the atoms do not match, the file structure is advanced to the
next atom (node) until a match with the starting query node is obtained. If
a match is obtained, the query proceeds to the next connected atom which is
compared with the next connected atom of the file structure. If these next
atoms do not match, the file structure is backtracked to the original
matching atom and another connected atom is selected for match. This
advancing/comparing/backtracking routine is continued until all atoms match
or all atom sequences of the query are exhausted. Overlap requires that all
atoms of the query match with all atoms of the file structure while
embedment requires that all of the atoms of the query be contained within
the file structure. A description of atom-by-atom matching is given in
Lynch, M. F.; Harrison, J. M.; Town, W. G.; Ash, J. E. Computer Handling of
Chemical Information, MacDonald, London and American Elsevier Inc., New
York, 1971 at pp. 73-74, all of which is herein incorporated by reference.
It is an object of this invention to extend both the overlap and
embedment matching concepts to Markush searching. Thus if the query SpMCN
IXa search is limited to overlap only, the query ISSR IXa1 will match only
with file ISSR XIIa2. If the matching criterion is relaxed to embedment,
ISSR XIIIa1 is also a valid match. It is not necessary to limit the
searching of a Markush query to a Markush file, e.g., the ISSRs of the
SpMCN representation also can be compared with both specific compound
representations such as X and XI and the ISSRs of the SpMCN representations
XIIa and XIIIa. At the overlap level of search, query IXa retrieves file
representations X and XIIa; at the embedment level of search, file
representations X, XI, XIIa, and XIIIa are retrieved. Single specific
compound queries also can be searched against the Markush file, e.g., VIII
matches with XIIa1 (overlap) and with XIIIa1 (embedment). Although not
illustrated, embedment and overlap criteria are also used at the generic
level of searching. Thus an implicit generic query representation,
alkyl-halide, overlaps an implicit generic file representation,
alkyl-halide, and is embedded in an implicit generic file representation,
carbocycle-alkyl-halide. Finally it is noted that the overlap criterion can
be applied to the entire SpMCN representation itself. Such a match
condition requires all structural elements of the file SpMCN be identical
to all structural elements of the query SpMCN, i.e., all ISSRs or the file
and query SpMCNs must be identical. Requiring the entire SpMCN
representation IXa to match at the overlap level permits only the retrieval
of file SpMCNs that are identical to it, i.e., contain both IXa1 and IXa2
but only those two implicit representations. For an entire SpMCN
representation to match at the embedment level, the file SpMCN
representation must contain all ISSRs of the query representation.
In order to convert SpMCN representations to GnMCN representations, it is
highly desirable to have a classification scheme that uses a small number
of controlled-vocabulary hierarchical terms that permit classification of
all groups of atoms likely to be encountered in a specific substance or a
Markush formulation. FIG. 6 illustrates such a classification scheme. The
overall structure of the classification scheme consists of breaking each
less-specific group into two mutually exclusive, more specific groups. The
general group "G" is used to handle groups of atoms that can not be easily
associated with a more specific group classification, e.g., an
electron-withdrawing group, a group containing nitrogen, etc. The G group
is classified further into two mutually exclusive groups: any cyclic group
(Cy) or any acyclic group (Ay). The cyclic group (Cy) is broken down into
any carbocycle group (Cb) or any heterocycle group (Hc). The carbocycle
group (Cb) characterizes any ring system containing only carbon atoms and
any attached hydrogen atoms. The Cb group may be attached to any other
group, including itself, or it may stand alone. The heterocycle group (Hc)
characterizes any ring system containing one or more hetero (noncarbon)
atoms and any attached hydrogen atoms. Similar to Cb, Hc may be attached to
any group, including Hc, or it may stand alone. A fused ring system, i.e.,
two or more rings joined at two or more atoms on each ring with each other,
is considered a single group while two rings joined to each other by an
acyclic bond is considered as two groups. Thus a naphthalene ring system is
designated as Cb while a biphenyl system would be characterized as Cb--Cb.
A quinoline ring system, which consists of a carbocycle fused to a
heterocycle, is considered as a single heterocycle group, Hc.
Moving to the acyclic side of the hierarchy, the acyclic group (Ay) is
broken down into any acyclic carbon (chain) group (Ch) or any acyclic
noncarbon (functional) group (Fg). The acyclic noncarbon group (Fg) is
further broken down into any acyclic noncarbon connecting group (Fc) or any
acyclic noncarbon terminal group (Ft). The terminal group (Ft) is
characterized as a single atom that is neither carbon or hydrogen but may
be attached to one or more hydrogens. The Ft group may stand alone
(unattached to any other group), e.g. NH sub 3, H sub 2 O, Cu, or it may be
attached to one and only one other group where the other group may be any
other group including Ft except that the Ft group cannot be bound to an
alkyl group (Ak) by a multiple bond since, by definition, an alkyl group
bound to a Ft group by a multiple bond is a Cg group. Thus C sub 6 H sub 5
--NH sub 2 transforms to Cb--Ft while an aldehyde such as CH sub 3 --CH
double bond O transforms to Cg double bond Ft and not Ak double bond Ft.
See infra Cg and Ak. The acyclic noncarbon connecting group (Fc) is defined
as a single atom that is neither carbon or hydrogen but may be attached to
one or more hydrogens and must be attached to two or more other groups
including itself, e.g., phenyl-O-phenyl is expressed as Cb--Fc--Cb. By
definition, Fc may not stand by itself or attached to only one other group.
The acyclic carbon group (Ch) is further broken down into an acyclic
carbon group (Cg) attached to an acyclic noncarbon terminal group (Ft) by a
multiple bond, or any other acyclic carbon group (Ak) not defined as Cg. By
definition, the Cg group can not stand alone. It must be attached to at
least one Ft group by a multiple bond and it may also be attached to other
groups, except Ak or Cg. The Ak group consists of a group of acyclic carbon
atoms and any attached hydrogen atoms that may stand alone or may be
attached to any group, except Cg or Ak. When a Cg is attached to an Ak or
another Cg or when an Ak is attached to a Cg or another Ak, the two groups
merge into the appropriate single group, e.g., Ft double bond Cg--Cg double
bond Ft becomes Ft double bond Cg double bond Ft, Ft double bond Cg--Ak
becomes Ft double bond Cg, and Ak--Ak becomes Ak. CH sub 3 --CH double bond
O becomes Cg double bond Ft; CH sub 3 --OH becomes Ak--Ft. The compound CH
sub 3 --CH sub 3 is not represented Ak--Ak but rather as simply Ak. The
compound O double bond CH--CH sub 2 CH sub 2 -- CH double bond O is not
represented as Ft double bond Cg--Ak--Ak--Cg double bond Ft but rather as
Ft double bond Cg double bond Ft. Table I illustrates the hierarchical
scheme using classification notation. FIG. 7 illustrates the use of the
hierarchy to transform SpMCN representations to GnMCN representations.
TABLE I
Hierarchical Generic Group Classification
G any chemical group
Cy any cyclic group
Cb any all carbon cyclic group
Hc any cyclic group containing at least one noncarbon atom
Ay any acyclic group
Ch any carbon acyclic group
Cg any carbon acyclic group attached by a multiple bond to
a terminal non-carbon acyclic group (Ft)
Ak any carbon acyclic group not defined by Cg
Fg any noncarbon acyclic group
Ft any noncarbon atom standing alone or attached to only one
other group
Fc any noncarbon atom attached to two or more other groups
The acyclic carbon group (Ak) can be further divided into two groups, one
group in which at least one multiple carbon-to-carbon bond must be present
and another group in which multiple carbon-to-carbon bonding is not
allowed. Alternatively, the Ak group can be divided on the basis of the
number of carbon atoms in the group. For a generalized file, separation
into two groups in which one group contains six or fewer carbon atoms and
another group containing more than six carbon atoms gives a relatively even
number of file compounds to the two groups. Other groups, such as Cb and Hc
can also be divided into more-specific, mutually-exclusive groups. E.g., Cb
could be divided into fused and non-fused groups or into groups based on
the number of rings within the group.
The above hierarchy is designed to meet the requirements of a search file
consisting of a broad range of chemical compounds likely to be found in a
broad definition of the chemical arts. However, this is not to imply that
this invention is limited to this hierarchy since it is anticipated that
other similar hierarchies may be defined to meet the needs of files limited
to particular kinds of compounds, e.g., inorganic compounds, cyclic
compounds, etc.
The classification hierarchy also is used to represent the generic
language used in the original Markush formulation so as to enable the
capture of this information in the SpMCN and GnMCN representations. For
example, as seen in structure XVIIIa of FIG. 8, Markush formulations may
contain terms that are in themselves generic, e.g., alkyl group, electron
with-drawing group, and heterocyclyl. Because the hierarchy defines all
possible chemical groups, generic terminology used in the Markush
formulation can be captured and used in the matching process. For example,
in going from structure XVIIIa to structure XVIIIb in FIG. 8, the alkyl
becomes Ak, the electron-withdrawing group becomes the general group G, and
the Heterocyclycl group becomes Hc. The appearance of generic groups from
two sources in the GnMCN (structure XVIIIc), i.e., from the Markush
formulation itself and from the real atoms in the SpMCN (structure XVIIIb)
requires additional processing considerations if high levels of recall are
to be obtained. To afford this additional processing, it is necessary to
distinguish generic groups generated from real atoms from those generated
from the original Markush formulation in the GnMCN representation. As shown
in GnMCN representation XVIIIc, generic groups generated from the Markush
formulation are designated with a prime (') to distinguish them fromgeneric groups generated from the real atoms shown in structure XVIIIb.
FIG. 8 also illustrates other aspects of a Markush formulation that are
anticipated by this invention. For example, the group R sub 2 is noted to
have no specific point of attachment in the Markush formulation (XVIIIa)
and in fact such formulation is intended to express the attachment of any
of the groups designated by R sub 2 in any of the open positions on the
six-membered ring. To handle such such a formulation, each of the R sub 2
groups is promulgated on each of the open nodes of the benzene ring in the
SpMCN representation (XVIIIb). The use of "n" to express an alkyl chain of
variable length in the Markush formulation (XVIIIa) is transformed to the
SpMCN representation (XVIIIb) by separate expression of each unique moiety
found in the "n" formulation. Finally it is noted that a Markush
formulation may contain "limiting" logic, i.e., "If R sub 1 =OMe, then R
sub 2 =Cl. Such "limiting" logic is associated and stored with the SpMCN
representation and is verified after query SpMCN and file SpMCN matching is
complete.
In order to apply known structure matching techniques used for single
compound matching to Markush searching in a straight-forward manner, it is
necessary to simplify complex SpMCN and GnMCN structures that result from
the representation of Markush formulations. Structure XIXb1 of FIG. 9
illustrates two such complexities, i.e., multiple ring formation when the
point of variability is within a ring itself and the formation of
"pseudo-rings" when the point of variability is within a chain structure.
By moving the point of variability to a node (atom) outside of the ring, it
is possible to eliminate such complex structural features (structures XIXb2
and XIXc). Occasionally, as shown in FIG. 10, the structural features of
the SpMCN may not permit the shifting of the point of variability outside
of the ring structure. This situation requires the use of a "null" or
"dummy" node, or, alternatively, the use of Boolean logic to represent the
various structural features. Representation XXb1 illustrates the use of the
null node (Nu) while representations XXb2-XXb4 illustrate the use of
individual representations linked by "OR" Boolean logic. The null node
representation or Boolean logic representations are also used with the
GnMCN representations (XXc1 and XXc2-XXc4, respectively).
One other feature that must be considered in the generation of the SpMCN
and GnMCN is the concept of variable hydrogen (H sub v). As was noted
earlier, it is a common practice to disregard the presence of hydrogen in
representing chemical structures, especially cyclic structures such as
representations Ia1-Ia4 of FIG. 1, and, in fact, such a convention often is
used in commercial structure-search systems. This is done typically to
minimize storage requirements. The hydrogen atoms attached to carbon atoms
are represented implicitly, i.e., the number of hydrogens attached to a
carbon atom can be determined from the number of bonds and the number of
non-hydrogen attachments. Although it is common notation to explicitly
represent hydrogen in acyclic structures, such specific representation
usually is not used in commercial structure search systems. Thus hydrogen
is typically not explicitly used in either cyclic or acyclic structure
search systems. See for example, "The CAS ONLINE Search System; General
System Design and Selection, Generation, and Use of Search Screens" by P.
G. Dittmar, N. A. Farmer, W. Fisanick, R. C. Haines, and J. Mockus (J.
Chem. Inf. Comput. Sci., 1983, 23, 93-102. ), all of which is hereby
incorporated by reference. However, in Markush formulations, hydrogen often
is used as one of the fragments in a variable group, and, in order to
achieve high recall, it is useful to express the hydrogen atom explicitly
in such cases. For example, in Markush formulation XXIa of FIG. 11, the
variable portion of the formulation contains the groups H, OH, and Cl. If
the H is not recorded in the SpMCN and GnMCN, the basic ISSR, XXIb1, is
"lost" to the implicit structure matching process at both the SpMCN and
GnMCN levels. To avoid such a loss, the variable group hydrogen is
explicitly represented in the SpMCN as H sub v so that the appropriate
ISSR, XXIb1, can be anticipated in the SpMCN structure matching process.
Likewise, variable hydrogen also is explicitly represented in the GnMCN
representation so that the appropriate basic IGSR can be anticipated in the
GnMCN structure matching process, e.g., ISSR, XXId1', of FIG. 11.
FIG. 11' illustrates the representation of H sub v and various
combinations of generic groups that are forbidden by the classification
scheme, e.g., Ak--Cg, Ak--Ak. In SpMCN representation XXIIb, H sub v is
explicitly expressed. On transformation to the GnMCN representation XXIIc,
several "forbidden" group combinations are observed, e.g., Ak--Cg, Ak--H
sub v, and Ak--Ak. These groups are appropriately combined according to the
hierarchical rules to give the IGSRs, Ak(XIId1'), Cg double bond
Ft(XXIId3'), and Ak--Fc--Ak(XXIId4'). Alternatively, the use of variable
hydrogen in the GnMCN can be avoided by merging it with the CH sub 2 to
give CH sub 3 and shifting the node of variability to the first carbon as
shown in SpMCN representation XXIIe. It should be noted that the node of
variability is shifted and not an individual substituent, i.e., if one
substituent is shifted then all substituents must be shifted. The shifting
of only one variable substituent of a set of variable substituents gives
rise to extraneous structures not implicit in the original Markush
formulation. SpMCN representation XXIIe is transformed to GnMCN XXIIf and
the forbidden group combinations eliminated in the IGSRs.
Because of the broad nature of the hierarchical classification used in
capturing the generic features of the original Markush formulation and in
transforming the SpMCN representation to the GnMCN representation,
considerable precision is lost in GnMCN structure matching if only the
hierarchical groups are used to capture these generic features. For
example, in FIG. 12, generic moieties a, b, and c and the specific moiety d
represent varying levels of atom and bond specificity that are lost if such
moieties are captured only as the generic group Hc. To improve GnMCN
matching precision, attributes of any generic expressions from the original
Markush formulation and real-atom moieties in the SpMCN are captured in an
attributes table and used in an additional matching step immediately
following the matching of a query and file group in query/file GnMCN
structure match. Alternatively, attribute matching can be applied as a
separate step following query/file GnMCN structure match for those cases
where a structure match has been found. As shown in FIG. 12, suitable
attributes for the Hc group are: (1) the identity of the hetero atom(s) and
their number, (2) the size of the ring, and (3) the type of bonding in the
ring. As shown in FIG. 12, moiety (a), heterocycle, conveys no additional
attributes not found in the Hc group itself; moiety (b), N-heterocycle,
conveys the fact that one or more of the heteroatoms (noncarbon atoms) in
the heterocycle are nitrogen atoms; moiety (c), a 6-membered heterocycle,
conveys the fact that there are six atoms in the cyclic structure, and (d),
a specific atom structure, conveys the identity and exact number of
heteroatoms (one nitrogen atom), the size of the ring (six atoms), and the
number and type of bonds found in the cyclic structure (six normalized
bonds). Attributes are derived from generic features described in the
original Markush (moieties (a)-(c)) as well as from the specific atom
structure found in the SpMCN representation (moiety (d)). The use of the
attribute table in the query/file comparison process is illustrated in
FIGS. 13 and 13' where the SpMCN, GnMCN, and Hc Attribute Table are given
for file representations XXIII-XXV and query representation XXVI. FIGS. 13
and 13' illustrate a set of file compounds in which the IGSR XXVIb has been
found to match at least one IGSR of each of the file representations
(XXIIIb-XXVb). When a match between the Hc group in the query and the Hc
group in the file representation is obtained during IGSR comparison, or
alternatively in the step after GnMCN comparison, the query Hc attributes
of the appropriate IGSR are compared with Hc attributes of the appropriate
file IGSR. As is evident, the query Hc attributes XXVIc are included only
within the file Hc attributes XXVc. Query attributes of the other groups
making up the query IGSR also are compared with attributes of the
corresponding groups of the file IGSRs. If all attributes of all groups of
at least one query IGSR are included within all of the corresponding groups
of at least one IGSR of a file GnMCN, the ISSRs of the associated query and
file SpMCN representations are compared in the next step. In comparing
attributes, only those attributes are compared for which a value exists in
both the query and file list. If a particular attribute in either the query
or file attribute list contains no information, no attempt is made to
compare that particular attribute. Thus in comparing the query attribute
list XXVIc with the file attribute list XXIVc, only attribute 1
(heteroatoms) was compared; no attempt was made to compare attributes 2 or
3. Depending on the precesion desired, additional detail can be used in
describing a particular attribute. For example, if the attribute is derived
from a real-atom configuration it is typically a closed set, i.e., limited
to a specific value, while an attribute derived from a generic term may be
an open set where additional or alternate values are possible. If attribute
1 of file representation XXIVc is considered as an alternate expression,
i.e., one or more nitrogen or one or more oxygen atoms, then the query
attributes XXVIc would also be included within its attributes. Such level
of detail in attribute description is typically handled by using the
appropriate Boolean logic.
FIGS. 2-4 illustrate the general search strategy of this invention, i.e.,
matching of IGSRs followed by matching of ISSRs. The IGSRs of FIGS. 3 and
3' and the ISSRs FIG. 4 were developed on the basis of an intuitive logic,
i.e., following from the Markush formulation itself. Although the
development of the ISSRs of FIG. 4 is aided significantly by the fact that
each ISSR must follow the usual connectivity rules of the chemical arts, at
the generic level, such connectivity rules are no longer available for the
development of IGSRs. One approach to the problem is to break the query
GnMCN into its simplest structural units and compare these units with the
file GnMCN representations. Thus in FIG. 3', query representation VIIc
breaks down into carbocycle-halide and carbocycle-alkyl. Comparing either
of these groups with the file GnMCN representations results in all but
structure IVc giving a match. At this point, two things are noted: (1) it
is not necessary to generate the implicit file GnMCN structures--the GnMCN
structure itself can be used to see if the fragment is contained within the
structure, and (2) by using a less specific implicit query representation,
an additional answer, irrelevant to be sure, is obtained, i.e.,
representation IIIc. Representation IIIc does, of course, drop from the
search at the SpMCN level of matching.
In order to obtain the highest level of specificity for the ISSRs and
IGSRs implicit in the respective SpMCN and GnMCN structures and to obviate
the need to expand each file and query SpMCN and GnMCN structure into its
respective ISSRs and IGSRs, it is highly desirable to have an indexing
system that allows for the recognition of all ISSR and IGSRs within the
respective query and file SpMCN and GnMCN structures. Such an indexing
system is illustrated in FIG. 14. Each atom and group is assigned a "flag"
(f sub n) that indicates the level of variability (zero level in the case
of atoms or groups in the base structure) and, if appropriate the k value
of the substituent, i.e., r sub k. As seen in Table II, such an indexing
system makes it possible to easily track all possible combinations of the
substituent groups.
One possible isolation scheme is to first select a fragment in a variable
group of the first level of variability, for example, Br (r(1)f(1)) and
then "prune" or discard other alternative fragments in the same variable
group and take all but one fragment in any other variable group as denoted
by the appropriate flags. Such a procedure leads directly to a single ISSR
or IGSR structure. By keeping track of which fragments have been used, all
ISSRs or IGSRs of the respective SpMCN or GnMCN representation can be
generated.
TABLE II
f(0) r(1)f(1) r(2)f(1) f(3)f(2)
ISSRs
Base structure
Br N(CH3)2 --
Base structure
Br NHCH2CH2 OCH3
Base structure
Br NHCH2CH2 SCH3
Base structure
Cl N(CH3)2 --
Base structure
Cl NHCH2CH2 OCH3
Base structure
Cl NHCH2CH2 SCH3
Base structure
OCH3 N(CH3)2 --
Base structure
OCH3 NHCH2CH2 OCH3
Base structure
OCH3 NHCH2CH2 SCH3
IGSRs
AkHc Ft FcAk2 --
AkHc Ft FcAk FcAk
AkHc Ft FcAk FcAk
AkHc Ft FcAk2 --
AkHc Ft FcAk FcAk
AkHc Ft FcAk FcAk
AkHc FcAk FcAk2 --
AkHc FcAk FcAk FcAk
AkHc FcAk FcAk FcAk
In its most basic form, such an index system requires that all r(k)
values be different. Thus in representation XXVIIc, the generation of an
implicit structure with two Ft groups attached to Hc is forbidden, since
such a structure would have two r(1) groups. The index system can also be
used to handle complex Markush formulations by using the node flags as a
way of breaking down a complex structure into separate structures linked by
appropriate Boolean logic. In such a treatment, it is necessary to capture
some atom and group detail from the previous and next higher level of
representation so as to preserve suitable detail for atom and group
matching techniques.
In capturing as much detail of the Markush formulation as possible,
generic features of the Markush formulation are captured in the SpMCN
representation as generic groups, e.g., representation XVIIIb of FIG. 8.
The use of such generic features in the SpMCN representation can result in
a matching failure if appropriate steps are not taken to allow for this
situation at the SpMCN matching level. To allow for appropriate generic
group matching at the SpMCN matching level, "roll-back" to the associated
GnMCN level of representation is used. The "roll-back" technique is
illustrated in FIG. 15. The query SpMCN representation XXVIIIa of FIG. 15
contains a phenyl group (uppermost portion of the representation) that is
implied in the generic group Cb of the file compound XXIXa (uppermost
portion of the representation).
Atom-by-atom structure matching will fail to match these two terms since
the query segment contains carbon atom (C) nodes which do not match with
the generic group (Cb) node. On identification of a mismatch due to a
comparison of a real atom node against a generic group node (as noted
earlier, generic groups in the SpMCN representation are distinguished by a
prime (') since they must originate from a generic feature in the Markushformulation; alternatively, such generic groups can be identified through a
table lookup), the appropriate real-atom part of the query (or file)
substance is rolled-back to its associated generic group node and the two
generic group nodes are compared. Thus, in the example, the real-atom
carbon atoms are rolled-back to the generic group node Cb and the Cb node
used to match the Cb node in the file representation. In the second segment
of the molecule, the query contains a generic feature (Ak) while the file
compound contains real atoms (CH sub 2 CH sub 2). In this case, the file
segment is rolled back to the generic group (Ak) and Ak used in the
matching process.
In using generic groups, one additional factor must be considered. For
example, in FIG. 15 the middle segment of the query (Ak) was successfully
compared with the rolled-backed Ak segment of the file group. If the query
segment had been a Ch group instead of an Ak group, a match would not have
been obtained even though, by hierarchical group definition, the Ak group
is included within the definition of a Ch group. To account for such
generic mismatch at either the GnMCN/GnMCN matching level itself or on
roll-back to the GnMCN group level in SpMCN/SpMCN matching, a particular
group within the hierarchy is expanded to include as alternate node values
all groups above it in the hierarchy (thus expanding Ch to include Ay and
G) and all lower lower groups in the hierarchy (Cg and Ak). By using such
an expansion for all file generic groups, a generic group node, such as Ch
or Ay, in the query will successfully match an expanded Ak group node in
the file substance and vice-versa.
Although the use of GnMCN representation matching and GnMCN attribute
matching reduces the amount of SpMCN representation matching considerably,
it is of further advantage to eliminate from the file as many irrelevant
GnMCN and SpMCN structures as possible prior to GnMCN group-by-group and
SpMCN atom-by-atom matching since these latter two processes are relatively
slow even when using high-speed computers. Such a reduction in file size is
accomplished by use of a modified version of known screening techniques
such as those described in "The CAS ONLINE Search System; General System
Design and Selection, Generation, and Use of Search Screens" by P. G.
Dittmar, N. A. Farmer, W. Fisanick, R. C. Haines, and J. Mockus (J. Chem.
Inf. Comput. Sci., 1983, 23, 93-102.), all of which is hereby incorporated
by reference. The generation of specific-atom screens, generic group
screens, and diagnostic screens is illustrated in FIG. 16. Specific atoms
screens are illustrated with representation XXXa of FIG. 16. Augmented Atom
(AA) screens are defined as screens involving an atom and its nearest
neighbor atoms and the attaching bonds. For example, in a first ISSR of
XXXa (ISSRs are not explicitly shown), the first carbon atom (C) is
attached to only one other carbon by a single bond (hydrogen atoms being
ignored). The second carbon is attached to the first carbon and a third
carbon; the third carbon is attached to the second carbon and an oxygen
(O); and the oxygen is attached to a carbon. In a second ISSR of XXXa, the
first carbon is attached to one other carbon, the second carbon is attached
to the first carbon and to a third carbon, the third carbon is attached to
the second carbon and to a nitrogen (N), and the nitrogen is attached to
the third carbon. In a third ISSR, the first carbon is attached to the
second carbon and the second carbon is attached to the first carbon and a
third carbon, and the third carbon is attached to the second carbon;
generic groups are specifically ignored in the generation of real atom
screens. The general notation for AA screens is to cite the atom under
consideration first followed by each attached atom in alphabetical order
with the bonding immediately preceding each attached atom.
Another type of screen is the Atom Sequence (AS) screen in which linear
sequences of four to six nonhydrogen atoms are described. For the ISSRs of
XXXa, there are two four-atom sequences, C--C--C--N and C--C--C--O. A third
type of screen is the Element Count (EC) screen which indicates the
presence of specific elements and for some of the more common elements, a
count of the number of atoms, e.g., six or more carbon atoms.
After generating the AA, EC, and AS screens, they are looked up in a
screen dictionary (Table III) to obtain their corresponding screens
numbers. The screen numbers correspond to specific bits in a bit string
illustrated in FIG. 16. It is noted that in screen fragment generation,
larger fragments are decomposed into smaller, more generic screen fragments
that are also included in the screen dictionary. For example, the AA screen
fragment, C--C--O (#4), is decomposed into C--C (#1) and C--O (#6). As
illustrated in FIG. 16, the screens for the three ISSRs of the SpMCN
representation XXXa turn on bits 1, 2, 4-7, 11-13, 17, and 19.
TABLE III
SCREEN DICTIONARY
SPECIFIC ATOM SCREENS
SCREEN NUMBER
EC SCREENS
C 11
O 13
N 12
Cl 14
AA SCREENS
C--C 1
Cl--C 9
N--C 7
O--C 6
C--C--C 2
C--C--Cl 10
C--C--N 5
C--C--O 4
C--C--C--C 3
C*C*C 8
C*C 15
AS SCREENS
C--C--C--C 18
C--C--C--Cl 20
C--C--C--N 19
C--C--C--O 17
(-- indicates a chain bond; * indicates a ring bond)
TABLE IV
SCREEN DICTIONARY
GENERIC SCREEN
SCREENS SCREEN NUMBER
Ak 58 (51,57,59)
Ak-Ay 37 (36,42,43,46,47,48; 51,57,58,59)
Ak-Cb 34 (33,36,43,46,48,49,50,60,61,62; 51,52,53,
57,58,59)
Ak-Cy 33 (36,43,46,48,49,50,60; 51,52,57,58,59)
Ak-Fg 32 (36,37,41,42,43,44,45,46,47,48; 51,56,57,58,59)
Ak-Ft 31 (32,36,37,38,39,40,41,42,43,44,45,46,47,48;
51,55,56,57,58,59)
Ak-G 36 (43,46,48; 51,57,58,59)
Ak-Hc 35 (33,36,43,46,48,49,50,60,63,64,65; 51,52,54,
57,58,59)
Ay 57 (51)
Ay-Ay 47 (46,48; 51,57)
Ay-Cb 62 (46,48,50,60; 51,52,53,57)
Ay-Ch 42 (43,46,47,48; 51,57,59)
Ay-Cy 50 (46,48,60; 51,52,57)
Ay-Fg 44 (45,46,47,48; 51,56,57)
Ay-Ft 39 (40,45,46,47,48; 51,55,56,57)
Ay-G 48 (46; 51,57)
Ay-Hc 64 (46,48,50,60,63; 51,52,54,57)
Cb 53 (51,52)
Cb-G 63 (46,60; 51,52,53)
Ch 59 (51,57)
Ch-Cb 61 (43,46,48,49,50,60,62; 51,52,53,57,59)
Ch-Cy 49 (43,46,48,50,60; 51,52,57,59)
Ch-Fg 41 (42,43,44,46,47,48; 51,56,57,59)
Ch-Ft 38 (39,40,41,42,43,44,46,47,48; 51,55,56,57,59)
Ch-G 43 (46,48; 51,57,59)
Ch-Hc 65 (43,46,48,49,50,60,63,64; 51,52,54,57,59)
Cy 52 (51)
Cy-G 60 (46; 51,52)
Fg 56 (51,57)
Fg-G 45 (46,48; 51,56,57)
Ft 55 (51,56,57)
Ft-G 40 (45,46,48; 51,55,56,57)
G 51
G-G 46 (51)
G-Hc 63 (46,60; 51,52,54)
Hc 54 (51,52)
TABLE V
SCREEN DICTIONARY
DIAGNOSTIC SCREENS
SCREENS SCREEN NUMBER
Ak' 78 (71, 77, 79)
Ay' 77 (71)
Cb' 73 (71, 72)
Ch' 79 (71, 77)
Cy' 72 (71)
Fg' 76 (71, 77)
Ft' 75 (71, 76, 77)
G' 71
Hc' 74 (71, 72)
For illustrative purposes, the screens and Specific-Atom Bit String used
to depict the ISSRs of SpMCN representation XXXa have been simplified
extensively. This is in no way intended to limit this invention to this
particular description; various other screens described in the article by
Dittmar, et. al, ibid, e.g., bond sequences, atom counts, etc. used for
specific-atom screening are also anticipated by this invention.
In developing screens for the GnMCN representation, it typically is not
necessary to generate as wide an array of screens as is used with the SpMCN
representation since many of the atomic features of the SpMCN
representation are compressed into a single group in the GnMCN
representation. Typically, AA doublets or dumbbells (a generic group joined
to one other generic group) and singlets (a generic group by itself) are
adequate for satisfactory screening at the generic group level. As with the
specific atoms screens, this description is given in simplified form to
teach the invention and is not intended to limit the number or types of
screens used for generic group screening.
The GnMCN representation XXXb contains two dumbbell screens, Ak--Ft and
Ak--Cy. Since Ak--Cy implies the more specific screens, Ak--Cb and Ak--Hc,
these screens must also be included as possible screens and are generated
at the time of bit string construction. If these two more specific screens
are not included, this particular file compound would be screened out when
in fact it is a potential match for a query such as Ak--Hc, i.e., the query
alkyl-heterocycle (Ak--Hc) is implicit in the file alkyl-cyclic group
(Ak--Cy). In examining the screen dictionary given in Table IV, it is
apparent that the Generic-Group screens, contain the more generic screens
for that screen. Thus the screen Ay--G includes the more generic doublet
term G--G and the singlets, Ay and G. As a result, the screen dictionary
automatically accounts for the matching of related, but more generic
fragments in which the more specific fragment is contained. For generic
representation XXXb, bits 31-33 and 36-65 are turned on in the
Generic-Group Bit String.
For generic groups that originate from generic expressions in the
original Markush formulation and, as noted in FIG. 8, denoted with a prime,
a special diagnostic bit screen is used. As with the generic screens, the
more generic terms are included with the term in the diagnostic screen
dictionary (Table V), e.g. Cy' includes G,; more specific terms are
generated at the time of bit string construction, e.g., Cb' and Hc' are
generated for Cy'. As a result, bits 71-74 are turned on in the Diagnostic
Bit String illustrated in FIG. 16.
In bit string generation for both generic and specific atom file
representations, screens for all possible ISSRs and IGSRs are included in a
single bit string. Moreover, the specific screens are generated without a
consideration of the boundaries which result when real atoms are
transformed into generic groups. Thus in representation XXXa of FIG. 16,
the AA screen C--C--O (No. 4) involved atoms (C--C) that transformed into
the Ak group and an atom (O) that transformed into the Ft group.
For query representations, each ISSR is considered as a separate entity
and real-atom screens are generated so as to restrict the real atoms
screens to atoms that transform to a single generic group or generic-group
combination. To simplify the illustration of the screening technique, the
real-atom screens are restricted to the boundaries of a single generic
group but this is not intended to restrict the scope of the invention.
Although real-atom screens must be restricted to the boundaries or the
corresponding generic counterpart, the generic counterpart may be defined
by any number of generic-group combinations. In representation XXXIa of
FIG. 17, the real atom screens are limited to those atoms that transform to
a single group, i.e., the screens are derived from three disconnected
fragments, CH sub 3 CH sub 2 CH sub 2, Cl, and OH. Only the AA screens
C--C--C and C--C are used for the group CH sub 3 CH sub 2. AA screens,
C--C--Cl and C--C--O, and AS screens, C--C--C--Cl and C--C--C--O, are not
used since these screens span two separate generic groups. Correspondingly,
the diagnostic generic screens are limited to the singlets, Ak and Ft.
Query XXXIa FIG. 17 along with the Specific-Atom and Generic Group Bit
Strings of FIG. 16 illustrate the screening technique in its basic form.
The IGSR Ak--Ft of XXXIb, corresponding to ISSR CH sub 3 CH sub 2 CH sub 2
Cl, gives rise to two disconnected singlets, Ak and Ft, which have
corresponding screen numbers 51, 55, 56, 57, 58, and 59. Using Boolean
logic, the query bit requirement is expressed as (Ak: 58 OR 51 OR 57 OR 59)
AND (Ft: 55 OR 51 OR 56 OR 57). (The Ak and Ft in the logic expression are
present for illustrative purposes only.) Examining the file Generic-Group
Bit String in FIG. 16, this initial condition is met. Proceeding to the
ISSRs, query ISSR CH sub 3 CH sub 2 CH sub 2 Cl, which is treated as two
disconnected fragments, CH sub 3 CH sub 2 CH sub 2 and Cl, requires that
the file Specific-Atom Bit String contain the following Boolean logic
query: 1 And 2 And 11 And 14. In examining the file Specific-Atom Bit
String of FIG. 16, it is apparent that this logic condition is not met
since bit 14 is off.
Proceeding to the next implicit structure (as is recalled only one
implicit structure of the query must match with an implicit structure of
the file representation), IGSR of CH sub 3 CH sub 2 CH sub 2 OH is the same
as that for CH sub 3 CH sub 2 CH sub 2 Cl, i.e., Ak--Ft, and thus it must
also match the Generic-Group Bit String of FIG. 16. The Boolean logic query
expression 1 And 2 And 11 And 13 is obtained for the ISSR, CH sub 3 CH sub
2 CH sub 2 OH, as the disconnected fragments CH sub 3 CH sub 2 CH sub 2 and
OH, and is found in the Specific-Atom Bit String of FIG. 16. Since both the
generic-group and specific-atom logic expressions of at least one IGSR and
corresponding ISSR of query XXXI is found in the file Generic-Group and
Specific-Atom Bit Strings, the associated GnMCN and SpMCN representations
are then compared using group-by-group and atom-by-atom techniques for each
IGSR and ISSR of both the file and query GnMCN and SpMCN representations.
Unlike group-by-group and atom-by-atom matching, which requires that each
IGSR and ISSR of both the file and query GnMCN and SpMCN representations be
compared, screening is satisfied as long as the generic-group
and
corresponding specific-atom screens of at least one IGSR and corresponding
ISSR is found in each aggregate file Generic-Group and corresponding
Specific-Atom Bit String, i.e., screens from all IGSRs and ISSRs for each
file GnMCN and and associated SpMCN representation.
ISSRs XXXIIa and XXXIIb of FIG. 17 illustrate the handling of a generic
feature found in the original Markush formulation. Look-up of the
individual generic groups of XXXIIb in the generic screen dictionary (Table
IV) results in the logic expression, (Ak: 58 OR 51 OR 57 OR 59) AND (Hc: 54
OR 51 OR 52) which is found in the Generic-Group Bit String of the file
representation of FIG. 16. In formulation a logic expression for an ISSR,
all generic groups are ignored. As a result, ISSR XXXIIa corresponds to the
fragment CH sub 3 CH sub 2 CH sub 2 and results in the logic expression, 1
AND 2 AND 11, which is also found in the Specific Bit String shown in FIG.
16. IGSR XXXIIb also illustrates the need to expand file representations to
include all more specific members, i.e., Ak--Cb and Ak--Hc in
representation XXXb. If the file representation had not been expanded to
include these members of the generic group, IGSR XXXIIb would have failed
to match.
ISSR XXXIIIa and IGSR XXXIIIb of FIG. 17 illustrates the use of the file
Diagnostic Bit String (FIG. 16) in handling a real-atom segment in the
query ISSR when the corresponding file ISSR has only a generic group
corresponding to that segment as a result of originating in a generic
feature of the original Markush for which no real-atom structure
information is available. The individual generic groups of IGSR XXXIIIb
give rise to the logic expression, (Ak: 58 OR 51 OR 57 OR 59) AND (Cb: 53
OR 51 OR 52) which is found in the Generic-Group Bit String of the file
representation XXXb. The specific atom groups of ISSR XXXIIIa give rise to
the logic expression, 1 AND 2 AND 8 AND 15. The logic expression does not
match with the file Specific-Atom Bit String of FIG. 16 since bits 8 and 15
of the file Bit String are off. However, it is noted that the file ISSR CH
sub 3 CH sub 2 CH sub 2 Cy does match the query ISSR. The failure to
recognize the match results from the fact that the query ISSR contains a
real-atom three-carbon ring-system from which the screen fragments C*C
(#15) and C*C*C (#8) are derived while the file ISSR contains a generic
group Cy from the original Markush formulation for which the screen
fragments C*C and C*C*C are not generated. In order to insure that the file
substance is not missed, the diagnostic screens exist in the file
Diagnostic Bit String to indicate when alternatives to real-atom query
screens must be used, i.e., they indicate when there are no real-atom
screens in the file Specific-Atom Bit String for a particular segment of
the file representation. (This is done since it is not practical to
generate and include in the Specific-Atom Bit String all possible real-atom
screens that correspond to the various generic groups.) For example, the
generic group corresponding to the query cyclic group in representation
XXXIIIa is Cb. On examining the diagnostic bit string for file
representation XXXa (FIG. 16), the diagnostic bit for Cb (#73) is "on". As
stated, the fact that the diagnostic screen is on for Cb means that the Cb
group originated from a generic feature of the Markush formulation and, as
such, no corresponding real atom screens are available in the file
Specific-Atom Bit String for this group. To obtain a match with this file
representation, it is necessary to use as an alternative to the real-atom
screens corresponding to Cb in the query logic expression the diagnostic
screens 73 OR 71 OR 72. As a result, screens 8 and 15 are not required to
be present for file representation XXXa if one of the above mentioned
diagnostic screens is present, and the resulting "reduced" real-atom screen
query logic expression for XXXIIIa is (1 AND 2 AND 11) AND ((8 AND 15) OR
(Cb': 73 OR 71 OR 72). Since this expression matches the Specific-Atom and
Diagnostic Bit Strings of FIG. 16, the file representation of FIG. 16
becomes a candidate for GnMCN group-by-group and SpMCN atom-by-atom
matching. As has been seen, the purpose of the file diagnostic screens is
to provide the necessary alternative to specific-atom screens in the query
logic expression when the file representation contains no real-atom screens
against which the specific-atom screens of the query expression can match.
The function of this special Boolean strategy is to insure as high a recall
as possible for the system.
For illustrative purposes, generic acid specific query/file screen
matching has been detailed in three steps, i.e., (1) matching of a generic
query logic expression with the generic file bit string (2) matching of a
specific query logic expression with a specific file bit string, and, (3)
using alternative diagnostic screens for specific-atom query screens when
the appropriate diagnostic screens are present in the file Diagnostic Bit
String. In practice, all three steps can be combined. A single specific and
generic logic expression is formulated for all segments of the ISSR and
within the specific logic expression diagnostic screens are used as
alternatives for the appropriate sets of real-atom screens that have been
derived from fragments corresponding to individual generic groups. Thus the
complete logic expression for implicit representations XXXIIIa and XXXIIIb
is: (((1 AND 2 AND 11) OR (Ak': 78 OR 71 OR 77 or 79)) AND (Ak: 58 OR 51 OR
57 OR 59)) AND ((1 AND 8 AND 15) OR (Cb' : 73 OR 71 OR 72) AND (Cb: 53 OR
51 OR 52)). This expression is passed against a combined Specific-Atom,
Generic-Group, and Diagnostic Bit String in one step with the appropriate
diagnostic screens being used according to their presence in the combined
file Bit String. Moreover, complete logic expressions for all ISSRs and
IGSRs in a single SpMCN and associated GnMCN query can be combined into a
single logic expression. Thus for SpMCN, XXXIa, and associated GnMCN,
XXXIb, the following logic expression is obtained: (((1 AND 1 AND 11) OR
(Ak': 78 OR 71 OR 77 OR 79)) AND (Ak: 58 OR 51 OR 57 OR 59)) AND (((13 OR
14) OR (Ft': 75 OR 71 OR 76 OR 77)) AND (Ft: 55 OR 51 OR 56 OR 57)).
In formulating specific atom screens, it is essential to limit specific
atom screens to the specific atoms defined by generic groups. As is
apparent, this specific atom screen limitation is necessary so that a
specific-atom group in a query can match a corresponding generic group in
the file that has no real-atom counterparts using the Boolean strategy
involving diagnostic screens. For illustrative purposes, the boundaries for
real-atom screens have been limited to to a corresponding single generic
group. In practice, combinations of generic groups are defined and all
real-atom screens within the generic-group combination used. For example,
the real-atom screens corresponding to the generic-group pair combination
Ak--Ft can be for the ISSR CH sub 3 CH sub 2 CH sub 2 Cl of representation
XXXIa of FIG. 17. In such a situation, the AA real-atom screen C--C--Cl
(#10) and AS screen C--C--C--Cl (#20) are used to eliminate additional
irrelevant structures from the file, i.e., to improve search precision.
However, in using such screens, the diagnostic screen boundaries must be
expanded by using the pair combination set Ak'--Ft, Ak--Ft', and Ak'--Ft',
i.e., Ak'--Ft OR Ak--Ft' OR Ak'--Ft' as alternatives for the real-atom
screen set. To afford even greater precision, both generic group pairs and
singlets can be used as diagnostic screens. For example, if the real-atom
screens derived from CH sub 3 CH sub 2 CH sub 2 Cl are denoted as "set a",
the screens from CH sub 3 CH sub 2 CH sub 2 as "set b", and the screens Cl
as "set c", and the diagnostic screens are Ak', Ft', and (Ak'--Ft OR
Ak--Ft' OR Ak'--Ft'); then the logic expression is ((set a OR (Ak'--Ft OR
Ak--Ft' OR Ak'--Ft') AND (set b OR Ak') AND (set c OR Ft')). The "set a"
screens are not used against a file substance with a GnMCN containing
Ak'--Ft, Ak--Ft', or Ak'--Ft', i.e., a GnMCN associated a SpMCN which has
an Ak' attached to a real-atom group corresponding to a Ft, a Ft' attached
to a real-atom group corresponding to an Ak, or a Ak'--Ft'.
Alternative to the above technique, an inverted file technique can be
used for the storage and searching of the file substance screens rather
then bit strings. In this technique, the screens (specific-atom, generic
group, and diagnostic) become file index terms and a common identifier for
the GnMCN and SpMCN representations associated with a given screen are
"inverted" under the file index term. The query screen logic expression is
developed in the same manner as for bit-string searching and is processed
against the screen index file giving as answers those substance identifiere
having the requisite query screens.
TABLE VI
INVERTED SCREEN FILE
SCREENS
Real Atom Generic Group Diagnostic
1 2 3 4 5 6 7 8 9 10
41
42
43
44
45
46
47
70
71
72
73
A A A A A A A A A A
B B B B B B B B
C C C C C C C C C
D D D D D D
E E E E E E E E E E E E E E
QUERY: ((2 AND 3) OR (71 OR 72)) AND (41 OR 43)
For example, in Table VI, "A" is the identifier for a file SpMCN and
associated GnMCN representations from which real-atom, generic-group, and
diagnostic screens are derived. "B" is the identifier for another file
SpMCN and associated GnMCN from which real-atom, generic-group, and
diagnostic screens are derived. "C," "D," and "E" are similar file
representation identifiers. Each file identifier is posted under the
corresponding screens generated from the SpMCN and GnMCN representation.
Appropriate file identifiers meeting the requisite Boolean logic conditions
of the query logic expression are retrieved as answers, here, file
identifiers "B" and "E." These GnMCN and SpMCN representations associated
with the answers are then further processed in group-by-group and
atom-by-atom searching.
FIG. 18 illustrates the general operation and computer hardware
facilities for the practice of this invention. A user terminal such as a
Hewlett-Packard 2647A (Hewlett-Packard Co., Cupertino, CA) is used for
graphic structure input. Query and search control uses the resources of a
large computer system such as one or more IBM 3081s (IBM Corp.,
Poughkeepsie, NY) and includes a telecommunications interfact and query and
search control procedures such as those described in Zeidner, C. R.; Amoss,
J. O.; Haines, R. C. "The CAS ONLINE architecture for Substructure
Searching" in "Proceedings of the 3rd National Online Meeting" Learned
Information, Inc.: Medford N.J., 1982; all of which is incorporated herein
by reference, software for handling graphic structure input such as the
Online Structure Input System (OLSIS: J. E. Blake, N. A. Farmer, and R. C.
Haines. "An interactive Computer Graphics System for Processing Chemical
Structure Diagrams." J. Chem. Inf. Comput. Sci., 1977, 17, 223-228), all of
which is herein incorporated by reference, a data-base management program
such as ADABAS (Software AG, Darmstadt, W. Germany) used during answer data
retrieval to access data bases such as patent files, abstracts files, etc.,
and output programs for online answers and off-line prints through a
high-speed printer such as the Xerox 9700 laser printer (not shown; Xerox
Corporation, El Sequndo, CA). Initial screen search is performed on a set
of minicomputers such as Digital Equipment Corp. (Maynard, MA.) PDP11/45s
while group-by-group and atom-by-atom searches and logic condition
verification are carried out on a set of minicomputers such as Digital
Equipment Corp. PDP 11/44s with one or more microcomputers such as the
Digital Equipment Corp. PDP 11/55 and PDP 11/44 used for system support,
maintenance, and backup.
In a typical search, a screen logic expression is formulated from the
query GnMCN and SpMCN structures and includes a special Boolean strategy to
ensure retrieval of all representations based on generic representations in
the original Markush formulation and is passed against the bit screen
files. Answers from the screening search are passed to generic
group-by-group search to determine if any IGSRs in the query GnMCN
representation are found in the file GnMCN representations. Attributes of
each generic-group node of the query representation are compared with the
attributes of each matching generic-group node of the file representations
during the GnMCN comparison. File answers from the generic-group/attribute
match are passed to an atom-by-atom search to determine if any ISSRs in the
query SpMCN representation are found in the file SpMCN representations. A
roll-back technique is used to achieve a high degree of recall during
atom-by-atom comparison by substituting a corresponding generic-group node
for a real-atom group of nodes when such a generic group node occurs in the
query or file substance. Resulting answers are checked to confirm that all
substituent logic conditions are met. The answers are returned to query
control where answers from appropriate data bases are retrieved and
forwarded to the user along with the substance search answers.
APPLICABILITY
This invention provides a powerful tool for Markush structure storage and
searching. It is designed to achieve total recall and very high precision
for the searching of Markush-type queries against a file of generic
substances of the type found in Markush patent claims. The query and file
representations handled range from specific structures to those with high
variability and generic features, including those that consist solely of
generic features. Both full-structure and substructure searching, at both
the specific-atom and generic-group level afford a wide range of search
capability.
A single multiple connectivity node structure for specific-atom
representations and an associated single multiple connectivity node
structure for generic-group representations afford a manageable file size
for the Markush formulations. An indexing system for for both the
specific-atom and generic-group multiple connectivity node structures
assures the generation and processing of all implict structures within the
generic-group and specific-atom multiple connectivity node structures as
well as providing for a means of dealing with highly complex structures.
A hierarchical generic-group classification scheme provides a means for
the input of generic features for both generic and file substances and,
more importantly, provides a controlled language for the fail-safe matching
of generic groups against generic groups and real-atom groups against
generic groups. This is accomplished by mapping all groups, both real-atom
and generic groups in the query and file substances, to a common set of
generic groups as defined by the hierarchy which can be dealt with by the
search routines, especially the initial screening step.
A screening technique is used to reduce the number of irrelevant file
answers for a particular query on a rapid and "fail-safe" basis. The
fail-safe result is achieved through a built-in handling of the generic vs.
generic and real-atom vs. generic matching via the common set of generic
groups as defined by the hierarchy. Real-atom screens are used to improve
precision in the screening process and, in order to preserve a high level
of recall, alternative Boolean diagnostic screens are provided in the query
screen logic expression for use when certain real-atom screens are not
appropriate.
High precision is achieved by matching file and query substances on a
group-by-group and atom-by-atom basis. This type of matching establishes
the required connectivity and arrangement of matching fragments lacking in
presently available systems. Attribute searching provides added precision
to the group-by-group by requiring that the attributes of two matching
groups also match. In atom-by-atom matching, a roll-back technique is used
to allow for the matching of a generic feature expressed as a generic group
in the query or file substance to match against a corresponding real-atom
group in the query or file substance.
While the forms of the invention herein disclosed constitute presently
preferred embodiments, many others are possible. It is not intended herein
to mention all of the possible equivalent forms or ramifications of the
invention. It is to be understood that the terms used herein are merely
descriptive rather than limiting, and that various changes may be made
without departing from the spirit or scope of the invention.
2/2,EM,SU/2 (Item 1 from file: 654)
DIALOG(R)File 654:US PAT.FULL.
(c) format only 2001 The Dialog Corp. All rts. reserv.
02999303
Utility
RELATIONAL DATABASE MANGEMENT SYSTEM FOR CHEMICAL STRUCTURE STORAGE,
SEARCHING AND RETRIEVAL
PATENT NO.: 5,950,192
ISSUED: September 07, 1999 (19990907)
INVENTOR(s): Moore, Jeffrey, Timonimun, MD (Maryland), US (United States of
America)
Brazil, Joanne, White Hall, MD (Maryland), US (United States
of America)
Hoover, Jeffrey R., Baltimore, MD (Maryland), US (United
States of America)
ASSIGNEE(s): Oxford Molecular Group, Inc , (A U.S. Company or Corporation),
Towson, MD (Maryland), US (United States of America)
APPL. NO.: 8-883,165
FILED: June 26, 1997 (19970626)
This application is a continuation, of application Ser. No. 08-715,708,
filed Sep. 19, 1996, now abandoned, which is a continuation application of
Ser. No. 08-288,503, filed Aug. 10, 1994, now U.S. Pat. No. 5,577,239, the
entire disclosure of which is incorporated herein by reference.
U.S. CLASS: 707-3 cross ref: 702-27
INTL CLASS: [6] G06F 17-30
FIELD OF SEARCH: 395-496; 395-497; 395-499; 395-600; 395-603; 707-3; 707-2;
707-1; 707-100; 707-102; 707-104; 707-22; 707-27; 707-19;
707-20; 702-22; 702-27; 702-19; 702-20
References Cited
U.S. PATENT DOCUMENTS
4,642,762 2/1987 Fisanick 707-3
4,811,217 3/1989 Tokizane et al. 364-300
4,855,931 8/1989 Saunders 364-499
5,025,388 6/1991 Cramer, III et al. 364-496
5,056,035 10/1991 Fujita 364-497
5,259,137 11/1993 Wilson et al. 364-496
5,367,058 11/1994 Pitner et al. 530-391.9
5,379,234 1/1995 Wilson et al. 364-496
5,386,507 1/1995 Teig et al. 395-161
5,418,944 5/1995 DiPace et al. 395-600
5,463,564 10/1995 Agrafiotis et al. 364-496
5,577,239 11/1996 Moore et al. 395-603
NON-U.S. PATENT DOCUMENTS
090 895 A2 10/1983 EP (European Patent Office)
213 483 A2 3/1987 EP (European Patent Office)
OTHER REFERENCES
Viking Instruments Corp. (Hewlett Packard); SpectraTrak Transportable GS/MS
Systems; (brochure)-No Date.
Chemical Structures, The International Language of Chemistry; Wendy A. War
(Ed.); "Interfacing DARC--Oracle" AJCM (Juus) de Jong (1988).
J. Chem. Inf. Comput. Sci. (1983) , vol. 23, No. 3; pp. 102-108; DARC
Substructure Search System: A New Approach to Chemical Information; Roger
Attias.
J. Chem. Inf. Comput. Sci. (1987), vol. 27, No. 2; pp. 74-82; DARC System:
Notions of Defined and Generic Substructures. Filiation and Coding of FREL
Substructure (SS) Classes; Jacques-Emile Dubois et al.
J. Chem. Inf. Comput. Sci. (1990), vol. 30, No. 2; pp. 191-199,
Substructure Search Systems. 1. Performance Comparison of the MACCS, DARC,
HTSS, and CAS Registry MVSSS, and S4 Substructure Search System; Martin G.
Hicks & Clemens.
J. Chem. Inf. Comput. sci. (1988), vol. 28, No. 4; pp. 221-226; An
Efficient Graph Approach to Matching Chemical Structures, O. Owolabi.
J. Chem. Inf. Comput. Sci. (1990), vol. 30, No. 4; pp. 332-339; Reactions
in the Beilstein Information System: Nonaporic Organic Synthesis; Martin G.
Hicks.
Analytica Chimica Acta, 235 (1990), pp. 87-92; Substructure Search Systems
for Large Chemical Data Bases; Martin G. Hicks et al.
J. Chem. Inf. Comput. Sci. (1991), vol. 31, No. 2; pp. 320-326; The
Beilstein Structure Registry System. 1. General Design; Laszio Domokos.
J. Chem. Inf. Comput. Sci. (1989), vol. 29, No. 4; pp. 255-260; 3DSearch; A
System for Three-Dimensional Substructure Searching; Robert P. Sheridan, et
al.
Substructure Searches of Chemical Structure Files; (Jan. 23, 1973);
Strategic Considerations in the Design of a Screening System for
Substructure Searches of Chemical Structure Files; George W. Adamson, et
al.
Chemical Structure Searching; (Jan. 21, 1975); An Efficient Design for
Chemical Structure Searching. I. The Screens; Alfred Feldman et al.
J. Chem. Inf. Comput. Sci. (1982), vol. No. 4; The Third BASIC Fragment
Search Dictionary; W. Graf, H. K. Kaindl, et al.
J. Chem. Inf. Comput. Sci. (1983), vol. 23, No. 3; The CAS Online Search
System. 1. General System Design and Selection, Generation, and Use of
Search Screens; P. G. Dittmar, et al.
Computer Chemical, ((1991), vol. 15, No. 2, pp. 103-107; A Central Atom
Based Algorithm and Computer Program for Substructure Search; Alf Dengler
and Ivar Ugi.
J. Chem. Inf. Comput. Sci. (1993), vol. 33, No. 4; pp. 545-547; Sturcture
Searching in Chemical Databases by Direct Lookup Methods; Baradley D.
Christie et al.
PRIMARY EXAMINER: Von Buhr, Maria N.
ATTORNEY, AGENT, OR FIRM: Dickstein Shapiro Morin & Oshinsky
CLAIMS: 14
EXEMPLARY CLAIM: 1
DRAWING PAGES: 7
DRAWING FIGURES: 12
ART UNIT: 277
FULL TEXT: 798 lines
FIELD OF THE INVENTION
The present invention relates to a relational database management system
that stores, searches and retrieves chemical structure information quickly
and easily.
BACKGROUND OF THE INVENTION
Chemical and pharmaceutical industries and chemical-related government
agencies commonly maintain large chemical substance databases. These
entities often provide structure-searching capabilities in association with
such databases. Recently, these organizations have been standardizing their
databases using relational database management systems (RDBMS) such as the
Oracle Relational Database Management System by Oracle Corporation, World
Headquarters, 500 Oracle Pkwy., Redwood Shores, Calif. 94065.
The advantages of integrating chemical structure information into an
RDBMS include: a closer integration with other related chemical data,
efficiency in both storage and retrieval of chemical structure data, and
better access to the chemical structure data by other related applications.
Unfortunately, chemical information systems have traditionally been built
using specialized database technology requiring, in many cases, hundreds of
thousands of lines of custom computer code. Systems of this type are often
both difficult to maintain, and difficult to adapt to changing hardware
technologies. These maintenance problems, coupled with a lack of
portability of these highly specialized systems, often lead to large
investments of time and money being allocated to relatively short-lived
systems.
The introduction of relational database technology provides an
opportunity to transfer a large amount of the database management
responsibility from the specialized database systems described above to a
standard widely-accepted technology. However, relational technology has
typically not been used as the basis for chemical information systems. This
is due to the fact that there are problems inherent in any attempt to cast
a chemical structure searching system problem into structured query
language (SQL)--the standard language of relational databases. These
problems include difficulty in storing and representing chemical structures
in a database. No chemical information system has yet been implemented
using only relational technology as its database component.
Several systems have attempted to achieve this goal but, as more fully
explained below, none have been able to develop a purely relational
database management system which is able to search and retrieve chemical
structure information easily and quickly.
For example, Molecular Access System (MACCS) and Integrated Scientific
Information System (ISIS) are both created by Molecular Design Ltd., MDL
Information Systems, 14600 Catalina Street, San Leandro, Calif. 94577.
These systems provide a stand-alone chemical information system wherein
chemical structures are stored as hierarchical structures. However, these
systems require large amounts of custom code, and are not maintained in a
relational database. Accordingly, they do not have the advantages of
relational technology listed above.
While it is true that these systems can be interfaced to a relational
database management system such as the Oracle Database Management System
noted above, it must be done using additional custom code and software that
converts hierarchical structures to the relational tables needed for such a
database. Therefore, it is difficult to incorporate the advantages of
relational technology into the MACCS and ISIS systems. Moreover, the
conversion software slows down overall performance speed.
In summary, these systems do not provide the advantages and capabilities
existing in the present invention.
The present invention overcomes the above-listed problems and
additionally has the following advantages:
(1) development and maintenance costs will be greatly reduced by using a
commercial database package. Accordingly, development efforts and benefits
can be more effectively directed toward aspects of system design, and
improvements in the underlying database technology will be automatically
transferred to the chemical information system. This shift of focus away
from database development concentrates the development and maintenance
efforts on improving the search strategy and the user interface, which are
the highly visible aspects of the system; (2) interfacing with other
information systems will be simplified since relational databases are
already used to store much of the non-structural chemical data used in
research and commercial settings; and (3) portability will be much less of
a design drawback since the amount of custom programming is minimal and can
easily be adapted to numerous types of technology. Therefore, the
portability responsibilities are mostly shouldered by the database
manufacturer itself, and not by the developer of the chemical storage
system.
SUMMARY OF THE INVENTION
The present invention overcomes the shortfalls in the art by developing a
chemical structure search system which expands the capabilities of existing
systems by capitalizing on the strengths of relational database technology.
The present invention allows the user to optimally store and search the
chemical structure information using various search strategies such as
multi-valued atoms, multi-typed bonds, Markush searching and various other
options in a relational database management system.
Furthermore, it provides a complete chemical information system which
includes modules for:
(1) exact structure searching;
(2) substructure searching;
(3) key searching;
(4) chemical name searching;
(5) molecular formula searching;
(6) registration of new molecules;
(7) structure import/export; and
(8) data editing.
Additionally, the present invention allows the routine integration of
chemical structure data with other related information such as inventory,
spectroscopic data and clinical data via standard relational database
methods to allow better usage of all types of chemical information in both
commercial and research settings.
By taking advantage of the data manipulation capabilities of relational
technology, this system will also introduce dynamic querying capabilities
which will allow the user to be notified of any new chemicals that are
entered into the database that are responsive to previously run queries.
This provides the functionality of relational views for chemical structure
information.
Additionally, structure classes can also be implemented which allow the
user to store certain types of information about particular types of
chemical structures such as steroids. Accordingly, users can later call up
this information in a quick and efficient manner without re-entering or
performing previously run queries.
With these and other objects, advantages and features of the invention
that may become apparent, the nature of the invention may be more clearly
understood by reference to the following detailed description of the
invention, the appended claims and the several drawings attached hereto.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT
The present invention makes use of standard relational database
technology such as that found in the commercial product Oracle which is
marketed by Oracle Corporation as noted above. All references to the
retrieval and storage of information will be done in a standard relational
database, and will use standard procedures for doing so, including
structured query language (SQL) commands. The operations and functions of
relational databases discussed in this patent application are well known to
those of ordinary skill in the database management field. Those operations
and functions can be found in numerous texts, including Oracle users' and
developers' manuals.
I. Hardware
Referring now to FIG. 9, the preferred embodiment of the relational
database management system for chemical structure storage and retrieval is
shown. A typical computer workstation 1 will contain a central processing
unit (CPU) 2, and main memory 3, and can be coupled to storage devices 4
such as magnetic disks, an input device such as a keyboard 5 or mouse, and
output device such as a computer monitor screen 6 and a printer 7. One or
more such storage devices may be utilized.
The preferred embodiment of the relational database management system for
storing, searching and retrieving chemical structure utilizes a
microprocessor, such as a Microvax 3100 model 900 operating with a VMS
5.5-2 operating system with at least two gigabytes of disk space and at
least 32 megabytes of RAM. The system can be provided with more memory to
speed up throughput access rates. The system could also be optionally
coupled to a local area network (LAN) or other communications
architecture/environment in order to link with other computer workstations
and have access to data from other systems.
II. Relational Database Interface
As noted above, one of the advantages of using relational databases for a
chemical structure search system is that there is no need on the part of
the developers to be concerned with portability, since the relational
database is a standard unto itself, that requires no special interface from
one type of system to the next.
In the present invention, the use of a standard relational database such
as the Oracle Relational Database Management System minimizes the
portability issues since they are available on virtually every platform.
Additionally, the present invention maintains the degree of portability
since it uses standard C with embedded SQL.
III. Registering New Structures
As shown in FIG. 6, to register a new structure in the database, the user
simply enters the atoms and bonds that make up the chemical structure by
typing the appropriate keys, or selecting the appropriate choices from
menus 22 in a standard chemical drawing software package such as Kekule,
marketed by PSI INTERNATIONAL, 810 Gleneagles Court, Suite 300, Towson, Md.
21286.
For each new structure that is registered (added) in the database,several steps must occur: (a) a connection table must be constructed and
stored in the database 24; (b) the system verifies that this structure is
not a duplicate 26; (c) at least one search key must be created and stored
in the database 28, and (d) information such as name, formula and registry
key number must be stored in the database 30. Each of these procedures will
be fully described below.
a. Construction of a Connection Table
For each structure to be registered in the database, a connection table
is constructed at step 24. This table stores information about each atom in
the structure including its atomic number, the identity of all of the
connected atoms, and the type of bond to each of these connected atoms. For
example, the connection table for a chemical structure to be added such as
structure S2 as depicted in FIG. 2 is shown in FIG. 3.
The table depicts the types of links that are stored between any two
given atoms in a structure. A single bond between two atoms is denoted by a
"1" while a double bond is denoted by a "2" and a triple bond is denoted by
a "3." This table is stored in a relational table along with its associated
registry number in a compressed sparse matrix form. The connection table
will be used for the Atom by Atom Matching (ABAM) process which is
described more fully below.
b. Search for Duplicates
The system then searches the existing structures to verify that no
duplicates exist in the database at step 26. If the structure has already
been entered into the system, it will not be entered again.
c. Creation of Search Keys
When a new structure of N atoms is registered in the system, it is
necessary to construct N search keys for the structure. These search keys
are stored as data in the relational database. Each search key results from
a unique numbering of the atoms with a different atom representing the
starting point of the key. In effect, N different search keys are stored
for each structure of N atoms.
To create effective search keys, it is necessary to derive an unambiguous
string of characters for each atom in the structure or query. The string is
a representation of the atomic environment of the starting atom. The
ordering of characters in the string cannot be sensitive to deletion of
portions of the structure or query. That is, deletion can remove portions
of the string (with subsequent replacement by wildcards in a query) but
cannot cause reordering of the remaining characters in the string. One such
algorithm, which builds the string by adding connectivity information using
a breadth-first graph traversal is detailed in FIG. 7, and is used in
subsequent examples.
As shown in FIG. 7, for each starting atom, the following process is used
to generate a search key. The process starts at step 40, and at step 41 all
atoms are marked as "unranked" and "unused". Next, at step 42 the starting
atom is marked as "used" and added to the key. Additionally, at step 43,
the starting atom is marked as the current atom.
So, for example, when reviewing structure S1 in FIG. 1, and starting with
the Bromine (Br) atom, the search key. The string would begin with "Br".
For clarity, the first code in the key is shown as the atomic symbol; in
practice, a one byte code is used for this purpose. Additionally, the
Bromine (Br) atom would be marked as "used" and set to the current atom.
At step 44, any unused neighbors are examined. In this example, the
Carbon (C sub 1) atom would be unused and accordingly, the system would
advance to step 45 where unused neighbors were ordered. Because there is
only one neighbor in this portion of the structure, and the ordering is not
terminated by an open site at step 46, and there is no open site at the
current atom at step 48, the system advances to step 49, where codes for
the neighbors, in order, are added to the key and marked as "used".
In this example, the letter "c" will be added to the key to indicate the
single bond to the Carbon (C sub 1) atom, and the Carbon (C sub 1) atom
will be marked as "used". The system next adds an end-of-atom marker to the
key at step 51. The key now reads "Br c .". The current atom (Br) is marked
as "ranked" at step 52.
The system next verifies that ordering was not terminated, and there was
no open site at step 53. The process continues at step 54 by examining if
any atoms in the key are unranked. In this example, the Carbon (C sub 1) is
unranked. Because the key is not too long (i.e., not longer than a
predefined length) at step 55, the Carbon (C sub 1) is chosen as the first
unranked atom in step 56. The process repeats itself starting at step 43
with the Carbon (C sub 1) atom as the current atom.
Now, the Carbon (C sub 1) is marked as the current atom, and the unused
neighbors are examined at step 44. Once again, there is only a single bond
to a Carbon (C sub 2) atom, and therefore the ordering at step 45 is
unnecessary. At step 49, a "c" is added to the key, and at step 51 the
end-of-atom marker is added to the key. Accordingly the key now reads "Br c
. c .".
This Carbon (C sub 1) atom is now marked as "ranked" at step 52. Once
again, the unranked atoms in the key are examined at step 54, and the first
unranked atom in the key is chosen at step 56. This Carbon (C sub 2) is now
set as the current atom at step 43, and the unused neighbors are examined
at step 44. There are two unused neighbors: a double bond to an Oxygen (O)
denoted by "e", and a single bond to a Carbon (C sub 3) denoted by "c". At
step 45, these bonds are ordered, with "c" taking precedence over "e".
These codes are then added to the key in order, and the atoms are marked as
"used" at step 49. Next, an end-of-atom marker is added to the key at step
51. The key now reads "Br c . c . c e .". The Carbon (C sub 2) atom is
marked as "ranked" , and the unranked atoms (the Oxygen and Carbon (C sub
3) atom) are examined at step 54.
At step 56, the first unranked atom in the key (C sub 3) is chosen, and
at step 43, it is set as the current atom. The process continues by
examining the single bond to the Carbon (C sub 4) emanating from this
Carbon (C sub 3), and accordingly "c" and "." are added to the key. The key
now reads "Br c . c . c e . c .".
The next unranked atom (Oxygen) is then set as the current atom at step
43. Given that there are no unused neighbors at step 44, an end-of-atom
marker is simply added to the key at step S1. The process once again
repeats with C sub 4 as the current atom. Because there are no unused
neighbors, "." is appended to the string, and the process stops for this
starting atom. The final search key would be: "Br c . c . c e . c . . .".
The same process is repeated for every starting atom in a given structure.
This process is also shown, step-by-step, in FIG. 12. (Steps 57, 58 and 59
in FIG. 7 will be explained in Section V. below.)
Utilizing this key generation algorithm (FIG. 7) with the illustrative
bonded atom codes shown in FIG. 4, the keys generated for structure S1
(FIG. 1) are:
1) Br c . c . c e . c . . .
2) C b c . . c e . c . . .
3) C c c e . b . c . . . .
4) O d . c c . b . c . . .
5) C c c . . c e . b . . .
6) C c . c . c e . b . . .
For structure S2 (FIG. 2), the keys generated are:
1) Cl c . c . c e . c . . .
2) C a c . . c e . c . . .
3) C c c e . a . c . . .
4) O d . c c . a . c . . .
5) C c c . . c e . a . . .
6) C c . c . c e . a . . .
These search keys are stored in the database with associated registry
numbers which correspond to the registry numbers of the connection tables
and associated information. Keys that are duplicated due to symmetry are
eliminated at registration time.
The steps required for processing a search are unaffected by the details
of the search key generation process. That is, any key generation process
which satisfies the conditions set forth in the opening paragraph of this
section (generation of an unambiguous character string for each atom in the
structure, etc.) can be utilized without modification of the search engine
software.
An additional process, which builds the string by listing structural
features found at each graph theoretical distance (level) from the starting
atom is detailed below.
As shown in FIG. 11, and using structure S1, the following process could
also be used to generate search strings. Beginning at step 80, all bonds in
a given structure are marked as "untraversed". A starting atom is chosen,
which in this case is the Bromine (Br) atom, at step 81, and added to the
key at step 82.
Because there are no open sites on this atom, the process continues at
step 85, where the system examines if any untraversed bonds to atoms at the
next level exist. In this case, there is an untraversed path to a Carbon (C
sub 1) atom. The system next determines that there is no open site at any
atom at this level at step 88 and, if not, continues at step 90, by
ordering all untraversed bonds to all atoms at the next level.
Because the wildcard flag has not been determined as having been set at
step 91, the system continues by adding the codes for the ordered paths to
the key at step 93. Then, at step 94, the system adds an end-of-level
marker to the key. Accordingly, the key now reads "Br c .", and all bonds
that are included in the key are marked as "traversed".
The system moves on to the next level with the Carbon (C sub 1) atom at
step 96. Once again, the system determines that there are untraversed paths
to atoms at the next level at step 85, and ultimately adds "c" and "." to
the key, at steps 93 and 94, respectively, after going through steps 88-92.
The system then marks these bonds as "traversed", and moves to the next
level beginning with the Carbon (C sub 2) atom. Once again, untraversed
paths are found at step 85, namely a double bond to an Oxygen (O) denoted
by "e", and a single bond to a Carbon (C sub 3) denoted by "c". These codes
are ordered at step 90, and are added to the key with "c" taking precedence
over "e" at step 93. Again, an end-of-level marker is added to the key at
step 94. The string now reads "Br c . c . c e .".
Next, the system advances to the next level in this structure, and
repeats the above process with the Carbon (C sub 3) atom. Once again "c "
and "." are added to the string.
Finally, the system moves to the next level at step 96 using the last
Carbon (C sub 4) atom. At step 85, the system determines that there are no
untraversed paths to atoms at the next level, and stops at step 87. The
final key reads: "Br c . c . c e . c ."
d. Associated Information Storage
Additional information about each structure can also be stored in the
database, such as registry key or other unique identifier of a structure,
name, and formula. The user may also define any additional information to
store and search using standard RDBMS technology.
IV. Implementation Issues
Each of the fragment codes comprising the search keys can be made to
occupy a single byte in the database. There are approximately 313 of these
fragment types existing in a large sample of structures. One byte allows
256 possibilities, three of which cannot be used. (Byte 0 cannot be used
due to its importance in programming, and two bytes used in the relational
database management system for its wildcard operation cannot be used, since
it is normally difficult to search these characters and use the SQL "Like"
operator in the same statement).
The remaining bytes can be divided into three groups: (1) those
representing the most common fragments, (2) those representing atoms whose
presence alone (regardless of bonding) is an effective screen, and (3)
those very rare atoms that can be grouped together; accordingly, every
search for these atoms is essentially a multi-valued search.
V. Processing of Query Substructure Searches
Performing a query against search keys is a relatively simple matter.
Each query structure generates one search key for each atom in the
structure that could be assigned an unambiguous fragment code. The search
keys of the query structure are generated by applying exactly the same
rules to the query structure as those used to generate the database search
key defined in III.b. above, with the only exception being treatment of
wildcards.
When a wildcard (i.e., a site in which no particular atom is necessary
for the search) is encountered, the process must either stop, or continue
the query by identifying all possibilities for the value of that wildcard.
Additionally, queries can easily accommodate multi-valued atoms and
multi-typed bonds or Markush searches in the same way that wildcards are
handled. The advantage of this methodology over standard screening
techniques is that it reflects the specificity of the query. For moderately
specific queries (i.e., few wild-cards), it enables remarkable selectivity
because of the length of the key.
As demonstrated with the generation of search keys above, the basis of
the search process is that the search keys are generated using one
unambiguous set of rules. So long as these rules are applied to the query
structure in exactly the same fashion, and since each database structure
has a key originating from each of its atoms, the results will be
standardized.
In order for the database structure to match the query structure, every
search key generated for the query must match one or more search keys
generated for the database structure. Additionally, if any query search key
fails to retrieve the structure, the query cannot be a substructure of that
structure. These rules make it possible to perform this extremely selective
screening process in a relational database with a single SELECT statement.
To generate a query, the user types in a structure in the same way that
new structures are entered, and can indicate where there are wildcards
(i.e., no particular atom necessary) and where there are multiple
acceptable types of atoms or bonds by indicating the specific atoms and
bonds that are acceptable in any given position.
As noted above, the query keys are generated in the same manner as the
search keys for the structures. Accordingly, any acceptable key generation
process may be used. Therefore, when generating a query key for Q1 in FIG.
5, and using the process shown in FIG. 7, the following steps occur.
The process begins at step 40, and at step 41, all atoms are marked as
"unranked" and "unused". A starting atom is chosen and marked as "used",
and the atom code is added to the key at step 42. In this example, the
Bromine (Br) atom will be the starting atom.
At step 43, Bromine (Br) is marked as the current atom, and at step 44,
it is determined that there are unused neighbors. The process continues at
step 45, with all unused neighbors being ordered.
Because there are no open sites at step 46, and because there is no open
site at the current atom at step 48, the codes for the neighbors are added
to the key in order and are marked as "used" at step 49. The process
continues with an end-of-atom marker being added to the key at step 51. The
key now currently reads "Br c .", and the Bromine (Br) atom is marked as
"ranked" at step 52.
Because the ordering was not terminated, and no open site was found at
the current atom at step 53, the process continues at step 54 by reviewing
the unranked atoms represented in the key. The system next verifies that a
maximum number of atoms has not been reached (the length of the query key
does not exceed a predetermined maximum length), and the first unranked
atom (C sub 1) is chosen at step 56.
The process continues at step 43 with the Carbon (C sub 1) atom being set
as a current atom. Again, all unused neighbors are examined at step 44, and
ordered at step 45. Because the system has still not encountered an open
site, the code for this bond (single bond to a Carbon) is added to the key
at step 49, with the end of atom marker added to the key at step 51. The
query key now reads "Br c . c .", and C sub 1 is marked as "ranked".
The process next repeats itself with C sub 2 as the first unranked atom
in the key at step 56. Accordingly, C sub 2 is marked as the current atom
at step 43, and at step 44 it is determined that there are still "unused"
neighbors. The unused neighbors are attempted to be ordered at step 45. The
bonds of the neighbors consist of a double bond to an Oxygen (O) denoted by
"e", and a wildcard (*).
The ordering is not terminated by the open site at step 46, because open
sites only terminate this process at step 46 if the open site does not
exist on the current atom. Because the open site exists on the current
atom, the "yes" branch is taken at step 481 and at step 50, codes for the
neighbors are added to the key with wildcard symbols around them. Next an
end-of-atom marker is added to the key at step 51. Accordingly, the current
string reads "Br c . c . % e %"; and the C sub 2 atom is marked as
"ranked."
At step 53, because an open site was found at the current atom, the
system advances to step 57. This string is a query key, so a wildcard is
added to the end of the key at step 58, and the process is stopped at step
59. Accordingly, the final string reads "Br c . c . % e % . %". The process
is then repeated with all other atoms as starting atoms.
Utilizing the key generation process (FIG. 7) with the illustrative
bonded atom codes as shown in FIG. 4, the keys generated for query Q1 (FIG.
5) are:
1) Br c . c . % e % . %
2) C b c . . % e % . %
3) C % c % e % . %
4) O d . % c % . %
The process illustrated in FIG. 11 can also be used to generate query
strings. The only notable addition would be that when an open site is
encountered at step 83 or 88, a wildcard flag is set at steps 84 or 89,
respectively. Then, from step 91, the system advances to step 92, and adds
wildcard (%) symbols before and after every code to be added to the string
at this step. Additionally, a wildcard (%) symbol is added to the end of
the string.
When a match exists between a query and a given structure, each of the
query keys will match one or more of the search keys of the structure.
Therefore, any one of the query keys may be used to retrieve a matching
structure. Statistical information is stored in the database allowing the
optimal query key to be used as the primary screen.
However, passing the screening phase alone is not enough to indicate that
a match has been found. The system must next verify that the query
structure is a subset of the structure by performing an atom by atom
matching (ABAM) process. To do this, a connection table is prepared for the
query structure in the same way that it is prepared for the newly
registered structures in III.a. above (see FIG. 3). These two connection
tables are then compared atom by atom, bond by bond. If every atom and bond
in the connection table for the query is found in the connection table for
the structure, the system returns a match.
Three cases are possible.
Case 1. The query is a substructure of the retrieval structure. Note that
Q1 (FIG. 5) is a substructure of S1 (FIG. 1). In this case, note that each
Q1 key matches an S1 key (Q1 keys 1, 2, 3, 4 match S1 keys 1, 2, 3, 4
respectively).
Case 2. The chosen key may retrieve a structure for which the query is
not a substructure. For example, Q1 key 4 would match S2 key 4 even though
Q1 is not a substructure of S2. However, the ABAM would eliminate this
structure.
Case 3. The chosen query key does not match any of the keys of a
particular structure. For example, Q1 key 1 does not match any of the keys
of S2. Therefore, S2 is eliminated as a match and ABAM is avoided.
Although one query key is typically used to drive the screening process,
the other query keys can be used a secondary screen. In Case 2 above, using
Q1 key 1 as a secondary screen would eliminate S2, thus avoiding ABAM.
The query, along with the matching results and other identifying
information (such as owner of query and name of query), are stored in the
relational database for later use and viewing.
The user may simply then advance through and view all of the structures
that are a match in the system which are displayed on a screen, such as
that shown in FIG. 10. If no structures match, the system will return a
message indicating this.
Due to the fact that most of the work is done during the initial
screening phase (comparing query strings to structure search keys), the
time-consuming atom by atom matching is done only on a relatively small
subset of the total structures in the database. Accordingly, this
methodology may be much quicker than other systems which perform the same
function.
Because all search queries and results are stored in the relational
database, the user, through standard relational database procedures, may
also list previously conducted searches, edit previously defined searches,
update or refresh previously run searches, view structures in any search,
and delete previous searches.
VI. Exact Structure (Identity) Searching
Identity searching involves finding a particular structure within a set
of database structures. This operation is performed by users, and is also
needed at the time of registration. Typically, this means finding a
structure in the database that matches the query exactly. An identity
search is a special case of the substructure search outlined above, but
having no open sites in the query.
Thus, the current substructure search method described above is an
adequate method for implementing identity searching, and as such will be
used accordingly. Additionally, the meaning of "exact match" can be
user-definable. The default definition limits the matching process to
element types and bonding. Users can also specify additional structural
information such as charge and mass values. This is performed at atom by
atom matching time.
VII. Chemical Name Searching
Chemical name searching has been a problem of special note in the field
of chemical information systems. Most chemical names are long and complex
strings which are not easily searchable by standard substring searching
mechanisms. This problem is compounded by the fact that most chemicals are
known by many systematic and/or tradenames.
Chemical name searching can be accomplished by storing and indexing
carefully defined name fragments, as well as indexing the complex strings
of the complete chemical names. Searching can be performed on a partial or
complete chemical name query using standard relational database technology.
To optimize the search, the query is degenerated into its constituent
chemical terms. The terms are sorted in ascending order by frequency of
occurrence found by looking up the number of compounds having a particular
term in a stored table. This stored table is created by scanning all names
of structures upon registration, and storing frequency information in that
table. Thus, this table acts as an index to chemical name fragments.
Given this list of chemical terms, the search can be performed by
intersecting the resulting SELECT statements or using one to drive a
correlated subquery.
Since the chemical name information is handled entirely by the relational
database, the data is then easily integrated with the rest of the chemical
information.
VIII. Molecular Formula Searching and Key Searching
Molecular formula can be done by using standard SQL string search methods
on all or part of the formula. Key searching (lookup by identifier) is a
standard SQL operation.
IX. Data Integration and Import/Export of Data
A significant advantage of basing a chemical information system in a
relational database is the ease with which the structure data can be
combined with related data, resulting in a complete, integrated system.
This allows information in other systems to be easily imported and exported
into the RDBMS using standard RDBMS functionality.
X. Dynamic Queries
As is true with all relational databases, the design of the system
decomposes into a series of entities, relationships, and functions. The
relationships among entities are rigorously defined since referential
integrity is the cornerstone of relational database design. A chemical
information system implemented using relational technology must be designed
with these considerations in mind.
A natural relationship exists between the database structures and the
resulting substructure searches with each search resulting in a set of
compound identifiers. In the present invention, a relational table is used
to store the set of identifiers during each search of the database. This is
implemented by creating a table to store general information about the
query (current user, date, query structure, options and search statistics).
A related table is created to store the identifiers of those structures
matching the query.
As new structures are registered in the system, the set of structures
identified as resulting from earlier queries becomes obsolete since the
structure database contains structures not present at the time of the
original search. This is, in a sense, a violation of referential integrity
because the relationship between structures and queries are not maintained.
In the present invention, however, the concept of dynamic queries is
introduced. When a new structure is registered, the system will examine
those queries designated as dynamic, and will add the identifier to the
search result set for each query matching the new structure. This process,
made simple by relational technology, allows the system to offer
functionality never before available in chemical information systems.
Dynamic queries are analogous to relational views. That is, they allow
searches of the database to be stored as objects in the database that are
always current. The following example will illustrate the type of
functionality made possible by dynamic queries.
As shown in FIG. 8, when a user performs a search at step 60
(substructure, chemical name, molecular formula), the system stores the
resulting set of compound identifiers at step 62. As the user examines
these compounds, the system flags each compound as having been viewed by
the user at step 64. Therefore, the system always knows the search results,
and the extent to which the user has reviewed them.
A user interested in a particular class of structures (e.g., steroids)
would perform a search once and designate the search as dynamic.
Thereafter, the search will be maintained automatically by the system. In
fact, the system would notify the user whenever a new steroid was
registered in the system at step 66. This is done by having the system
perform all dynamic queries on any newly registered molecule as it is
registered into the database at step 68 and notifying the user if a match
occurs at step 70. The user could then view the previously unseen results
at step 72 without having to repeat the query or view previously seen
results.
While dynamic queries would not have much importance with relatively
static databases, they would have many uses in the system serving a
research environment. Heavy use of dynamic queries could require allocation
of significant amounts of disk space for storing search results.
Additionally, the performance of the registration process could be
adversely affected by the presence of a large number of dynamic queries.
These potential problems can be controlled by the introduction of a
resource allocation system with each user being assigned two quotas. The
first quota controls the number of dynamic queries that the user can have
active at any one time, which will protect performance at the time of the
registration process. The second quota will control the total number of
structure identifiers that each user has stored by dynamic queries to
conserve disk space.
Alternatively, users would be allowed to disable these quota systems, but
this may slow the system during the registration process, or may exhaust
disk space for the database.
XI. Structure Classes
The division of chemical structures into classes based on overlapping
criteria (e.g., functional groups, ring systems) have long been used as an
organizational technique in chemistry. Chemical information systems
typically provide for this class system by allowing users to intersect the
results of different searches. While this intersection is a necessary
feature of any chemical information system, it does not address the
fundamental importance of the classification
schemes used in chemistry. In
the present invention, a mechanism will be provided for maintaining any
number of classification schemes in the database for structures. These
schemes or structure classes can be privately defined by individual users,
or can be used as a system-wide search aid.
A "structure class" is defined to be a set of structure identifiers
resulting from a substructure search, a chemical name search, a molecular
formula search, or by a combination of these searches. Structure classes
are an application of dynamic queries used to limit the scope of future
searches. For example, the system may maintain a structure class for
steroids. When a user performs a search, he or she can designate that the
result should be restricted to the members of the steroid class.
Accordingly, the user could simply query the database for all steroids that
have a particular substructure without drawing the entire steroid ring.
This results in two primary benefits: The first benefit is that queries
are simplified, i.e., there is no need to draw complex queries, and the
second benefit is that the screening phase need only be applied to those
compounds already known to be members of the structure class.
Dynamic queries and structure classes both exhibit a common benefit--the
overhead involved in structure searching is encountered only once (when the
dynamic query or structure class is defined) and additional overhead is
distributed evenly across subsequent updates to the chemical structure
database.
From the preceding description, it is evident that the invention has been
described in detail by reference to a particular embodiment adapted for use
in the field of chemistry. Although this invention offers many advantages
in this field, it may be used in other fields wherein structure data is
stored advantageously as well. Accordingly, this invention is not intended
to be limited by the details of the preferred embodiment described above,
but rather by the terms of the appended claims.
2/2,EM,SU/3 (Item 2 from file: 654)
DIALOG(R)File 654:US PAT.FULL.
(c) format only 2001 The Dialog Corp. All rts. reserv.
02592280
Utility
CHEMICAL STRUCTURE STORAGE, SEARCHING AND RETRIEVAL SYSTEM
PATENT NO.: 5,577,239
ISSUED: November 19, 1996 (19961119)
INVENTOR(s): Moore, Jeffrey, 12 Breezy Tree Ct., Timonimun, MD (Maryland),
US (United States of America), 21093
Brazil, Joanne, 4500 Jolly Acres Rd., White Hall, MD
(Maryland), US (United States of America), 21161
Hoover, Jeffrey R., 8639 Willow Oak Rd., Baltimore, MD
(Maryland), US (United States of America), 21234
[Assignee Code(s): 68000]
EXTRA INFO: Assignment transaction [Reassigned], recorded October 12,
1994 (19941012)
Assignment transaction [Reassigned], recorded January 21,
1997 (19970121)
APPL. NO.: 8-288,503
FILED: August 10, 1994 (19940810)
U.S. CLASS: 707-3 cross ref: 702-27
INTL CLASS: [6] G06F 17-30
FIELD OF SEARCH: 364-DIG.1; 364-DIG.2; 364-496; 364-497; 364-499; 395-600
References Cited
U.S. PATENT DOCUMENTS
4,642,762 2/1987 Fisanick 364-300
4,811,217 3/1989 Tokizane et al. 364-300
4,855,931 8/1989 Saunders 364-499
5,025,388 6/1991 Cramer, III et al. 364-496
5,056,035 10/1991 Fujita 364-497
5,249,137 2/1993 Wilson et al. 364-496
5,367,058 11/1994 Pitner et al. 530-391.9
5,379,234 1/1995 Wilson et al. 364-496
5,386,507 1/1995 Teig et al. 395-161
5,418,944 5/1995 DiPace et al. 395-600
5,463,564 10/1995 Agrafiotis et al. 364-496
OTHER REFERENCES
Viking Instruments Corp. (Hewlett Packard); Spectra Trak Transportable
GC/MS System; (brochure), No date.
Chemical Structure, The International Language of Chemistry; Wendy A. War
(Ed.); "Interfacing DARC-Oracle" AJCM (Juus) de Jong (1988).
J. Chem. Inf. Comput. Sci. (1983), vol. 23, No. 3 pp. 102-108; DARC
Substructure Search System; A New Approach to Chemical Information; Roger
Attias.
J. Chem. Inf. Comput. Sci. (1987), vol. 27, No. 2; pp. 74-82; DARC System;
Notions of Defined and Generic Substructures. Filiation and Coding of FREL
Substructure (SS) Classes; Jacques-Emile Dubois et al.
J. Chem. Inf. Comput. Sci. (1990), vol. 30, No. 2; pp. 191-199,
Substructure Search Systems, 1, Performance Comparison of the MACCS, DARC,
HTSS, CAS Registry MVSSS, and S4 Substructure Search Systems; Martin G.
Hicks.
J. Chem. Inf. Comput. Sci. (1988), vol. 28, No. 4; pp. 221-226; An
Efficient Graph Approach to Matching Chemical Structures, O. Owolabi.
J. Chem. Inf. Comput. Sci. (1990), vol. 30, No. 4; pp. 332-339; Reactions
in the Bellstein Information System: Nonaporic Organic Synthesis; Martin G.
Hicks.
Analytica Chimica Acta, 235 (1990), pp. 87-92; Substructure Search Systems
for Large Chemical Data Bases; Martin G. Hicks et al.
J. Chem. Inf. Comput. Sci. (1991), vol. 31, No. 2; pp. 320-326; The
Bellstein Structure Registry System, 1, General Design; Laszio Domokos.
J. Chem. Inf. Comput. Sci. (1989), vol. 29, No. 4; pp. 255-260; 3DSearch; A
System for Three-Dimensional Substructure Searching; Robert P. Sheridan, et
al.
Substructure Searches of Chemical Structure Files; (Jan. 23, 1973);
Strategic Considerations in the Design of a Screening System for
Substructure Searches of Chemical Structure Files; George W. Adamson, et
al.
Chemical Structure Searching; (Jan. 21, 1975); An Efficient Design for
Chemical Structure Searching, I, The Screens; Alfred Feldman et al.
J. Chem. Inf. Comput. Sci. (1982), vol. 22, No. 4; The Third BASIC Fragment
Search Dictionary; W. Graf, H. K. Kaindl, et al.
J. Chem. Inf. Comput. Sci. (1983), vol. 23, No. 3; The CAS ONLINE Search
System, 1, General System Design and Selection, Generation, and Use of
Search Screens; P. G. Dittmar, et al.
Computer Chemical, (1991), vol. 15, No. 2; pp. 103-107; A Central Atom
Based Algorithm and Computer Program for Substructure Search; Alf Dengler
and Ivar Ugi.
J. Chem. Inf. Comput. Sci. (1993), vol. 33, No. 4; pp. 545-547; Structure
Searching in Chemical Databases by Direct Lookup Methods; Bradley D.
Christie et al.
J. Chem. Inf. Comput. Sci. (1993); vol. 33, No. 4; pp. 539-541;
Substructure Searching on Very Large Files by Using Multiple Storage
Techniques; Alexander Bartmann et al.
PRIMARY EXAMINER: Black, Thomas G.
ASST. EXAMINER: Von Buhr, Maria N.
ATTORNEY, AGENT, OR FIRM: Dickstein Shapiro Morin & Oshinsky LLP
CLAIMS: 12
EXEMPLARY CLAIM: 1
DRAWING PAGES: 7
DRAWING FIGURES: 12
ART UNIT: 237
FULL TEXT: 791 lines
FIELD OF THE INVENTION
The present invention relates to a relational database management system
that stores, searches and retrieves chemical structure information quickly
and easily.
BACKGROUND OF THE INVENTION
Chemical and pharmaceutical industries and chemical-related government
agencies commonly maintain large chemical substance databases. These
entities often provide structure-searching capabilities in association with
such databases. Recently, these organizations have been standardizing their
databases using relational database management systems (RDBMS) such as the
Oracle Relational Database Management System by Oracle Corporation, World
Headquarters, 500 Oracle Pkwy., Redwood Shores, Calif. 94065.
The advantages of integrating chemical structure information into an
RDBMS include: a closer integration with other related chemical data,
efficiency in both storage and retrieval of chemical structure data, and
better access to the chemical structure data by other related applications.
Unfortunately, chemical information systems have traditionally been built
using specialized database technology requiring, in many cases, hundreds of
thousands of lines of custom computer code. Systems of this type are often
both difficult to maintain, and difficult to adapt to changing hardware
technologies. These maintenance problems, coupled with a lack of
portability of these highly specialized systems, often lead to large
investments of time and money being allocated to relatively short-lived
systems.
The introduction of relational database technology provides an
opportunity to transfer a large amount of the database management
responsibility from the specialized database systems described above to a
standard widely-accepted technology. However, relational technology has
typically not been used as the basis for chemical information systems. This
is due to the fact that there are problems inherent in any attempt to cast
a chemical structure searching system problem into structured query
language (SQL)--the standard language of relational databases. These
problems include difficulty in storing and representing chemical structures
in a database. No chemical information system has yet been implemented
using only relational technology as its database component.
Several systems have attempted to achieve this goal but, as more fully
explained below, none have been able to develop a purely relational
database management system which is able to search and retrieve chemical
structure information easily and quickly.
For example, Molecular Access System (MACCS) and Integrated Scientific
Information System (ISIS) are both created by Molecular Design Ltd., MDL
Information Systems, 14600 Catalina Street, San Leandro, Calif. 94577.
These systems provide a stand-alone chemical information system wherein
chemical structures are stored as hierarchical structures. However, these
systems require large amounts of custom code, and are not maintained in a
relational database. Accordingly, they do not have the advantages of
relational technology listed above.
While it is true that these systems can be interfaced to a relational
database management system such as the Oracle Database Management System
noted above, it must be done using additional custom code and software that
converts hierarchical structures to the relational tables needed for such a
database. Therefore, it is difficult to incorporate the advantages of
relational technology into the MACCS and ISIS systems. Moreover, the
conversion software slows down overall performance speed.
In summary, these systems do not provide the advantages and capabilities
existing in the present invention.
The present invention overcomes the above-listed problems and
additionally has the following advantages: (1) development and maintenance
costs will be greatly reduced by using a commercial database package.
Accordingly, development efforts and benefits can be more effectively
directed toward aspects of system design, and improvements in the
underlying database technology will be automatically transferred to the
chemical information system. This shift of focus away from database
development concentrates the development and maintenance efforts on
improving the search strategy and the user interface, which are the highly
visible aspects of the system; (2) interfacing with other information
systems will be simplified since relational databases are already used to
store much of the non-structural chemical data used in research and
commercial settings; and (3) portability will be much less of a design
drawback since the amount of custom programming is minimal and can easily
be adapted to numerous types of technology. Therefore, the portability
responsibilities are mostly shouldered by the database manufacturer itself,
and not by the developer of the chemical storage system.
SUMMARY OF THE INVENTION
The present invention overcomes the shortfalls in the art by developing a
chemical structure search system which expands the capabilities of existing
systems by capitalizing on the strengths of relational database technology.
The present invention allows the user to optimally store and search the
chemical structure information using various search strategies such as
multi-valued atoms, multi-typed bonds, Markush searching and various other
options in a relational database management system.
Furthermore, it provides a complete chemical information system which
includes modules for:
(1) exact structure searching;
(2) substructure searching;
(3) key searching;
(4) chemical name searching;
(5) molecular formula searching;
(6) registration of new molecules;
(7) structure import/export; and
(8) data editing.
Additionally, the present invention allows the routine integration of
chemical structure data with other related information such as inventory,
spectroscopic data and clinical data via standard relational database
methods to allow better usage of all types of chemical information in both
commercial and research settings.
By taking advantage of the data manipulation capabilities of relational
technology, this system will also introduce dynamic querying capabilities
which will allow the user to be notified of any new chemicals that are
entered into the database that are responsive to previously run queries.
This provides the functionality of relational views for chemical structure
information.
Additionally, structure classes can also be implemented which allow the
user to store certain types of information about particular types of
chemical structures such as steroids. Accordingly, users can later call up
this information in a quick and efficient manner without re-entering or
performing previously run queries.
With these and other objects, advantages and features of the invention
that may become apparent, the nature of the invention may be more clearly
understood by reference to the following detailed description of the
invention, the appended claims and the several drawings attached hereto.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT
The present invention makes use of standard relational database
technology such as that found in the commercial product Oracle which is
marketed by Oracle Corporation as noted above. All references to the
retrieval and storage of information will be done in a standard relational
database, and will use standard procedures for doing so, including
structured query language (SQL) commands. The operations and functions of
relational databases discussed in this patent application are well known to
those of ordinary skill in the database management field. Those operations
and functions can be found in numerous texts, including Oracle users' and
developers' manuals.
I. HARDWARE
Referring now to FIG. 9, the preferred embodiment of the relational
database management system for chemical structure storage and retrieval is
shown. A typical computer workstation 1 will contain a central processing
unit (CPU) 2, and main memory 3, and can be coupled to storage devices 4
such as magnetic disks, an input device such as a keyboard 5 or mouse, and
output device such as a computer monitor screen 6 and a printer 7. One or
more such storage devices may be utilized.
The preferred embodiment of the relational database management system for
storing, searching and retrieving chemical structure utilizes a
microprocessor, such as a Microvax 3100 model 900 operating with a VMS
5.5-2 operating system with at least two gigabytes of disk space and at
least 32 megabytes of RAM. The system can be provided with more memory to
speed up throughput access rates. The system could also be optionally
coupled to a local area network (LAN) or other communications
architecture/environment in order to link with other computer workstations
and have access to data from other systems.
II. RELATIONAL DATABASE INTERFACE
As noted above, one of the advantages of using relational databases for a
chemical structure search system is that there is no need on the part of
the developers to be concerned with portability, since the relational
database is a standard unto itself, that requires no special interface from
one type of system to the next.
In the present invention, the use of a standard relational database such
as the Oracle Relational Database Management System minimizes the
portability issues since they are available on virtually every platform.
Additionally, the present invention maintains the degree of portability
since it uses standard C with embedded SQL.
III. REGISTERING NEW STRUCTURES
As shown in FIG. 6, to register a new structure in the database, the user
simply enters the atoms and bonds that make up the chemical structure by
typing the appropriate keys, or selecting the appropriate choices from
menus 22 in a standard chemical drawing software package such as Kekule,
marketed by PSI INTERNATIONAL, 810 Gleneagles Court, Suite 300, Towson, Md.
21286.
For each new structure that is registered (added) in the database,
several steps must occur: (a) a connection table must be constructed and
stored in the database 24; (b) the system verifies that this structure is
not a duplicate 26; (c) at least one search key must be created and stored
in the database 28, and (d) information such as name, formula and registry
key number must be stored in the database 30. Each or, these procedures
will be fully described below.
a. Construction of a Connection Table
For each structure to be registered in the database, a connection table
is constructed at step 24. This table stores information about each atom in
the structure including its atomic number, the identity of all of the
connected atoms, and the type of bond to each of these connected atoms. For
example, the connection table for a chemical structure to be added such as
structure S2 as depicted in FIG. 2 is shown in FIG. 3.
The table depicts the types of links that are stored between any two
given atoms in a structure. A single bond between two atoms is denoted by a
"1" while a double bond is denoted by a "2" and a triple bond is denoted by
a "3." This table is stored in a relational table along with its associated
registry number in a compressed sparse matrix form. The connection table
will be used for the Atom by Atom Matching (ABAM) process which is
described more fully below.
b. Search for Duplicates
The system then searches the existing structures to verify that no
duplicates exist in the database at step 26. If the structure has already
been entered into the system, it will not be entered again.
c. Creation of Search Keys
When a new structure of N atoms is registered in the system, it is
necessary to construct N search keys for the structure. These search keys
are stored as data in the relational database. Each search key results from
a unique numbering of the atoms with a different atom representing the
starting point of the key. In effect, N different search keys are stored
for each structure of N atoms.
To create effective search keys, it is necessary to derive an unambiguous
string of characters for each atom in the structure or query. The string is
a representation of the atomic environment of the starting atom. The
ordering of characters in the string cannot be sensitive to deletion of
portions of the structure or query. That is, deletion can remove portions
of the string (with subsequent replacement by wildcards in a query) but
cannot cause reordering of the remaining characters in the string. One such
algorithm, which builds the string by adding connectivity information using
a breadth-first graph traversal is detailed in FIG. 7, and is used in
subsequent examples.
As shown in FIG. 7, for each starting atom, the following process is used
to generate a search key. The process starts at step 40, and at step 41 all
atoms are marked as "unranked" and "unused". Next, at step 42 the starting
atom is marked as "used" and added to the key. Additionally, at step 43,
the starting atom is marked as the current atom.
So, for example, when reviewing structure S1 in FIG. 1, and starting with
the Bromine (Br) atom, the search string would begin with "Br". For
clarity, the first code in the key is shown as the atomic symbol; in
practice, a one byte code is used for this purpose. Additionally, the
Bromine (Br) atom would be marked as "used" and set to the current atom.
At step 44, any unused neighbors are examined. In this example, the
Carbon (C) atom would be unused and accordingly, the system would advance
to step 45 where unused neighbors were ordered. Because there is only one
neighbor in this portion of the structure, and the ordering is not
terminated by an open site at step 46, and there is no open site at the
current atom at step 48, the system advances to step 49, where codes for
the neighbors, in order, are added to the key and marked as "used".
In this example, the letter "c" will be added to the key to indicate the
single bond to the Carbon (C sub 1) atom, and the Carbon (C sub 1) atom
will be marked as "used". The system next adds an end-of-atom marker to the
key at step 51. The key now reads "Br c .". The current atom (Br) is marked
as "ranked" at step 52.
The system next verifies that ordering was not terminated, and there was
no open site at step 53. The process continues at step 54 by examining if
any atoms in the key are unranked. In this example, the Carbon (C sub 1) is
unranked. Because the key is not too long (i.e., not longer than a
predefined length) at step 55, the Carbon (C sub 1) is chosen as the first
unranked atom in step 56. The process repeats itself starting at step 43
with the Carbon (C sub 1) atom as the current atom.
Now, the Carbon (C sub 1) is marked as the current atom, and the unused
neighbors are examined at step 44. Once again, there is only a single bond
to a Carbon (C sub 2) atom, and therefore the ordering at step 45 is
unnecessary. At step 49, a "c" is added to the key, and at step 51 the
end-of-atom marker is added to the key. Accordingly the key now reads "Br c
. c .".
This Carbon (C sub 1) atom is now marked as "ranked" at step 52. Once
again, the unranked atoms in the key are examined at step 54, and the first
unranked atom in the key is chosen at step 56. This Carbon (C sub 2) is now
set as the current atom at step 43, and the unused neighbors are examined
at step 44. There are two unused neighbors: a double bond to an Oxygen (O)
denoted by "e", and a single bond to a Carbon (C sub 3) denoted by "c". At
step 45, these bonds are ordered, with "c" taking precedence over "e".
These codes are then added to the key in order, and the atoms are marked as
"used" at step 49. Next, an end-of-atom marker is added to the key at step
51. The key now reads "Br c . c . c e .". The Carbon (C sub 2) atom is
marked as "ranked", and the unranked atoms (the Oxygen and Carbon (C sub 3)
atom) are examined at step 54.
At step 56, the first unranked atom in the key (C sub 3) is chosen, and
at step 43, it is set as the current atom. The process continues by
examining the single bond to the Carbon (C sub 4) emanating from this
Carbon (C sub 3), and accordingly "c" and "." are added to the key. The key
now reads "Br c . c . c e . c .".
The next unranked atom (Oxygen) is then set as the current atom at step
43. Given that there are no unused neighbors at step 44, an end-of-atom
marker is simply added to the key at step S1. The process once again
repeats with C sub 4 as the current atom. Because there are no unused
neighbors, "." is appended to the string, and the process stops for this
starting atom. The final search key would be: "Br c . c . c e . c . . .".
The same process is repeated for every starting atom in a given structure.
This process is also shown, step-by-step, in FIG. 12. (Steps 57, 58 and 59
in FIG. 7 will be explained in Section V. below.)
Utilizing this key generation algorithm (FIG. 7) with the illustrative
bonded atom codes shown in FIG. 4, the keys generated for structure S1
(FIG. 1) are:
1) Br c . c . c e . c . . .
2) C b c . . c e . c . . .
3) C c c e . b . c . . . .
4) O d .c c . b . c . . .
5) C c c . . c e . b . . .
6) C c . c . c e . b . . .
For structure S2 (FIG. 2), the keys generated are:
1) Cl c . c . c e . c . . .
2) C a c . . c e . c . . .
3) C c c e . a . c . . . .
4) O d . c c . a . c . . .
5) C c c . . c e . a . . .
6) C c . c . c e . a . . .
These search keys are stored in the database with associated registry
numbers which correspond to the registry numbers of the connection tables
and associated information. Keys that are duplicated due to symmetry are
eliminated at registration time.
The steps required for processing a search are unaffected by the details
of the search key generation process. That is, any key generation process
which satisfies the conditions set forth in the opening paragraph of this
section (generation of an unambiguous character string for each atom in the
structure, etc.) can be utilized without modification of the search engine
software.
An additional process, which builds the string by listing structural
features found at each graph theoretical distance (level) from the starting
atom is detailed below.
As shown in FIG. 11, and using structure S1, the following process could
also be used to generate search strings. Beginning at step 80, all bonds in
a given structure are marked as "untraversed". A starting atom is chosen,
which in this case is the Bromine (Br) atom, at step 81, and added to the
key at step 82.
Because there are no open sites on this atom, the process continues at
step 85, where the system examines if any untraversed bonds to atoms at the
next level exist. In this case, there is an untraversed path to a Carbon (C
sub 1) atom. The system next determines that there is no open site at any
atom at this level at step 88 and, if not, continues at step 90, by
ordering all untraversed bonds to all atoms at the next level.
Because the wildcard flag has not been determined as having been set at
step 91, the system continues by adding the codes for the ordered paths to
the key at step 93. Then, at step 94, the system adds an end-of-level
marker to the key Accordingly, the key now reads "Br c ." and all bonds
that are included in the key are marked as "traversed".
The system moves on to the next level with the Carbon (C sub 1) atom at
step 96. Once again, the system determines that there are untraversed paths
to atoms at the next level at step 85, and ultimately adds "c" and "." to
the key, at steps 93 and 94, respectively, after going through steps 88-92.
The system then marks these bonds as "traversed", and moves to the next
level beginning with the Carbon (C sub 2) atom. Once again, untraversed
paths are found at step 85, namely a double bond to an Oxygen (O) denoted
by "e", and a single bond to a Carbon (C sub 3) denoted by "c". These codes
are ordered at step 90, and are added to the key with "c" taking precedence
over "e" at step 93. Again, an end-of-level marker is added to the key at
step 94. The string now reads "Br c . c . c e .".
Next, the system advances to the next level in this structure, and
repeats the above process with the Carbon (C sub 3) atom. Once again "c"
and "." are added to the string.
Finally, the system moves to the next level at step 96 using the last
Carbon (C sub 4) atom. At step 85, the system determines that there are no
untraversed paths to atoms at the next level, and stops at step 87. The
final key reads: "Br c . c . c e . c . ".
d. Associated Information Storage
Additional information about each structure can also be stored in the
database, such as registry key or other unique identifier of a structure,
name, and formula. The user may also define any additional information to
store and search using standard RDBMS technology.
IV. IMPLEMENTATION ISSUES
Each of the fragment codes comprising the search keys can be made to
occupy a single byte in the database. There are approximately 313 of these
fragment types existing in a large sample of structures. One byte allows
256 possibilities, three of which cannot be used. (Byte 0 cannot be used
due to its importance in programming, and two bytes used in the relational
database management system for its wildcard operation cannot be used, since
it is normally difficult to search these characters and use the SQL "Like"
operator in the same statement).
The remaining bytes can be divided into three groups: (1) those
representing the most common fragments, (2) those representing atoms whose
presence alone (regardless of bonding) is an effective screen, and (3)
those very rare atoms that can be grouped together; accordingly, every
search for these atoms is essentially a multi-valued search.
V. PROCESSING OF QUERY SUBSTRUCTURE SEARCHES
Performing a query against search keys is a relatively simple matter.
Each query structure generates one search key for each atom in the
structure that could be assigned an unambiguous fragment code. The search
keys of the query structure are generated by applying exactly the same
rules to the query structure as those used to generate the database search
key defined in III.b. above, with the only exception being treatment of
wildcards.
When a wildcard (i.e., a site in which no particular atom is necessary
for the search) is encountered, the process must either stop, or continue
the query by identifying all possibilities for the value of that wildcard.
Additionally, queries can easily accommodate multi-valued atoms and
multi-typed bonds or Markush searches in the same way that wildcards are
handled. The advantage of this methodology over standard screening
techniques is that it reflects the specificity of the query. For moderately
specific queries (i.e., few wild-cards), it enables remarkable selectivity
because of the length of the key.
As demonstrated with the generation of search keys above, the basis of
the search process is that the search keys are generated using one
unambiguous set of rules. So long as these rules are applied to the query
structure in exactly the same fashion, and since each database structure
has a key originating from each of its atoms, the results will be
standardized.
In order for the database structure to match the query structure, every
search key generated for the query must match one or more search keys
generated for the database structure. Additionally, if any query search key
fails to retrieve the structure, the query cannot be a substructure of that
structure. These rules make it possible to perform this extremely selective
screening process in a relational database with a single SELECT statement.
To generate a query, the user types in a structure in the same way that
new structures are entered, and can indicate where there are wildcards
(i.e., no particular atom necessary) and where there are multiple
acceptable types of atoms or bonds by indicating the specific atoms and
bonds that are acceptable in any given position.
As noted above, the query keys are generated in the same manner as the
search keys for the structures. Accordingly, any acceptable key generation
process may be used. Therefore, when generating a query key for Q1 in FIG.
5, and using the process shown in FIG. 7, the following steps occur.
The process begins at step 40, and at step 41, all atoms are marked as
"unranked" and "unused". A starting atom is chosen and marked as "used",
and the atom code is added to the key at step 42. In this example, the
Bromine (Br) atom will be the starting atom.
At step 43, Bromine (Br) is marked as the current atom, and at step 44,
it is determined that there are unused neighbors. The process continues at
step 45, with all unused neighbors being ordered.
Because there are no open sites at step 46, and because there is no open
site at the current atom at step 48, the codes for the neighbors are added
to the key in order and are marked as "used" at step 49. The process
continues with an end-of-atom marker being added to the key at step 51 The
key now currently reads "Br c .", and the Bromine (Br) atom is marked as
"ranked" at step 52.
Because the ordering was not terminated, and no open site was found at
the current atom at step 53, the process continues at step 54 by reviewing
the unranked atoms represented in the key. The system next verifies that a
maximum number of atoms has not been reached (the length of the query key
does not exceed a predetermined maximum length), and the first unranked
atom (C sub 1) is chosen at step 56.
The process continues at step 43 with the Carbon (C sub 1) atom being set
as a current atom. Again, all unused neighbors are examined at step 44, and
ordered at step 45. Because the system has still not encountered an open
site, the code for this bond (single bond to a Carbon) is added to the key
at step 49, with the end of atom marker added to the key at step 51 The
query key now reads "Br c . c .", and C sub 1 is marked as "ranked".
The process next repeats itself with C sub 2 as the first unranked atom
in the key at step 56. Accordingly, C sub 2 is marked as the current atom
at step 43, and at step 44 it is determined that there are still "unused"
neighbors. The unused neighbors are attempted to be ordered at step 45. The
bonds of the neighbors consist of a double bond to an Oxygen (O) denoted by
"e", and a wildcard (*).
The ordering is not terminated by the open site at step 46, because open
sites only terminate this process at step 46 if the open site does not
exist on the current atom. Because the open site exists on the current
atom, the "yes" branch is taken at step 481 and at step 50, codes for the
neighbors are added to the key with wildcard symbols around them. Next an
end-of-atom marker is added to the key at step 51. Accordingly, the current
string reads "Br c . c . % e %"; and the C sub 2 atom is marked as
"ranked."
At step 53, because an open site was found at the current atom, the
system advances to step 57. This string is a query key, so a wildcard is
added to the end of the key at step 58, and the process is stopped at step
59. Accordingly, the final string reads "Br c . c . % e % . %". The process
is then repeated with all other atoms as starting atoms.
Utilizing the key generation process (FIG. 7) with the illustrative
bonded atom codes as shown in FIG. 4, the keys generated for query Q1 (FIG.
5) are:
1) Br c . c . % e % . %
2) C b c . . % e % . %
3) C % c % e % . %
4) O d . % c % . %
The process illustrated in FIG. 11 can also be used to generate query
strings. The only notable addition would be that when an open site is
encountered at step 83 or 88, a wildcard flag is set at steps 84 or 89,
respectively. Then, from step 91, the system advances to step 92, and adds
wildcard (%) symbols before and after every code to be added to the string
at this step. Additionally, a wildcard (%) symbol is added to the end of
the string.
When a match exists between a query and a given structure, each of the
query keys will match one or more of the search keys of the structure.
Therefore, any one of the query keys may be used to retrieve a matching
structure. Statistical information is stored in the database allowing the
optimal query key to be used as the primary screen.
However, passing the screening phase alone is not enough to indicate that
a match has been found. The system must next verify that the query
structure is a subset of the structure by performing an atom by atom
matching (ABAM) process. To do this, a connection table is prepared for the
query structure in the same way that it is prepared for the newly
registered structures in III.a. above (see FIG. 3). These two connection
tables are then compared atom by atom, bond by bond. If every atom and bond
in the connection table for the query is found in the connection table for
the structure, the system returns a match.
Three cases are possible.
Case 1. The query is a substructure of the retrieval structure. Note that
Q1 (FIG. 5) is a substructure of S1 (FIG. 1). In this case, note that each
Q1 key matches an S1 key (Q1 keys 1, 2, 3, 4 match S1 keys 1, 2, 3, 4
respectively).
Case 2. The chosen key may retrieve a structure for which the query is
not a substructure. For example, Q1 key 4 would match S2 key 4 even though
Q1 is not a substructure of S2. However, the ABAM would eliminate this
structure.
Case 3. The chosen query key does not match any of the keys of a
particular structure. For example, Q1 key 1 does not match any of the keys
of S2. Therefore, S2 is eliminated as a match and ABAM is avoided.
Although one query key is typically used to drive the screening process,
the other query keys can be used a secondary screen. In Case 2 above, using
Q1 key 1 as a secondary screen would eliminate S2, thus avoiding ABAM.
The query, along with the matching results and other identifying
information (such as owner of query and name of query), are stored in the
relational database for later use and viewing.
The user may simply then advance through and view all of the structures
that are a match in the system which are displayed on a screen, such as
that shown in FIG. 10. If no structures match, the system will return a
message indicating this.
Due to the fact that most of the work is done during the initial
screening phase (comparing query strings to structure search keys), the
time-consuming atom by atom matching is done only on a relatively small
subset of the total structures in the database. Accordingly, this
methodology may be much quicker than other systems which perform the same
function.
Because all search queries and results are stored in the relational
database, the user, through standard relational database procedures, may
also list previously conducted searches, edit previously defined searches,
update or refresh previously run searches, view structures in any search,
and delete previous searches.
VI. EXACT STRUCTURE (IDENTITY) SEARCHING
Identity searching involves finding a particular structure within a set
of database structures. This operation is performed by users, and is also
needed at the time of registration. Typically, this means finding a
structure in the database that matches the query exactly. An identity
search is a special case of the substructure search outlined above, but
having no open sites in the query.
Thus, the current substructure search method described above is an
adequate method for implementing identity searching, and as such will be
used accordingly. Additionally, the meaning of "exact match" can be
user-definable. The default definition limits the matching process to
element types and bonding. Users can also specify additional structural
information such as charge and mass values. This is performed at atom by
atom matching time.
VII. CHEMICAL NAME SEARCHING
Chemical name searching has been a problem of special note in the field
of chemical information systems. Most chemical names are long and complex
strings which are not easily searchable by standard substring searching
mechanisms. This problem is compounded by the fact that most chemicals are
known by many systematic and/or tradenames.
Chemical name searching can be accomplished by storing and indexing
carefully defined name fragments, as well as indexing the complex strings
of the complete chemical names. Searching can be performed on a partial or
complete chemical name query using standard relational database technology.
To optimize the search, the query is degenerated into its constituent
chemical terms. The terms are sorted in ascending order by frequency of
occurrence found by looking up the number of compounds having a particular
term in a stored table. This stored table is created by scanning all names
of structures upon registration, and storing frequency information in that
table. Thus, this table acts as an index to chemical name fragments.
Given this list of chemical terms, the search can be performed by
intersecting the resulting SELECT statements or using one to drive a
correlated subquery.
Since the chemical name information is handled entirely by the relational
database, the data is then easily integrated with the rest of the chemical
information.
VIII. MOLECULAR FORMULA SEARCHING AND KEY SEARCHING
Molecular formula can be done by using standard SQL string search methods
on all or part of the formula. Key searching (lookup by identifier) is a
standard SQL operation.
IX. DATA INTEGRATION AND IMPORT/EXPORT OF DATA
A significant advantage of basing a chemical information system in a
relational database is the ease with which the structure data can be
combined with related data, resulting in a complete, integrated system.
This allows information in other systems to be easily imported and exported
into the RDBMS using standard RDBMS functionality.
X. DYNAMIC QUERIES
As is true with all relational databases, the design of the system
decomposes into a series of entities, relationships, and functions. The
relationships among entities are rigorously defined since referential
integrity is the cornerstone of relational database design. A chemical
information system implemented using relational technology must be designed
with these considerations in mind.
A natural relationship exists between the database structures and the
resulting substructure searches with each search resulting in a set of
compound identifiers. In the present invention, a relational table is used
to store the set of identifiers during each search of the database. This is
implemented by creating a table to store general information about the
query (current user, date, query structure, options and search statistics).
A related table is created to store the identifiers of those structures
matching the query.
As new structures are registered in the system, the set of structures
identified as resulting from earlier queries becomes obsolete since the
structure database contains structures not present at the time of the
original search. This is, in a sense, a violation of referential integrity
because the relationship between structures and queries are not maintained.
In the present invention, however, the concept of dynamic queries is
introduced. When a new structure is registered, the system will examine
those queries designated as dynamic, and will add the identifier to the
search result set for each query matching the new structure. This process,
made simple by relational technology, allows the system to offer
functionality never before available in chemical information systems.
Dynamic queries are analogous to relational views. That is, they allow
searches of the database to be stored as objects in the database that are
always current. The following example will illustrate the type of
functionality made possible by dynamic queries.
As shown in FIG. 8, when a user performs a search at step 60
(substructure, chemical name, molecular formula), the system stores the
resulting set of compound identifiers at step 62. As the user examines
these compounds, the system flags each compound as having been viewed by
the user at step 64. Therefore, the system always knows the search results,
and the extent to which the user has reviewed them.
A user interested in a particular class of structures (e.g., steroids)
would perform a search once and designate the search as dynamic.
Thereafter, the search will be maintained automatically by the system. In
fact, the system would notify the user whenever a new steroid was
registered in the system at step 66. This is done by having the system
perform all dynamic queries on any newly registered molecule as it is
registered into the database at step 68 and notifying the user if a match
occurs at step 70. The user could then view the previously unseen results
at step 72 without having to repeat the query or view previously seen
results.
While dynamic queries would not have much importance with relatively
static databases, they would have many uses in the system serving a
research environment. Heavy use of dynamic queries could require allocation
of significant amounts of disk space for storing search results.
Additionally, the performance of the registration process could be
adversely affected by the presence of a large number of dynamic queries.
These potential problems can be controlled by the introduction of a
resource allocation system with each user being assigned two quotas. The
first quota controls the number of dynamic queries that the user can have
active at any one time, which will protect performance at the time of the
registration process. The second quota will control the total number of
structure identifiers that each user has stored by dynamic queries to
conserve disk space.
Alternatively, ushers would be allowed to disable these quota systems,
but this may slow the system during the registration process, or may
exhaust disk space for the database.
XI. STRUCTURE CLASSES
The division of chemical structures into classes based on overlapping
criteria (e.g., functional groups, ring systems) have long been used as an
organizational technique in chemistry. Chemical information systems
typically provide for this class system by allowing users to intersect the
results of different searches. While this intersection is a necessary
feature of any chemical information system, it does not address the
fundamental importance of the classification schemes used in chemistry. In
the present invention, a mechanism will be provided for maintaining any
number of classification schemes in the database for structures. These
schemes or structure classes can be privately defined by individual users,
or can be used as a system-wide search aid.
A "structure class" is defined to be a set of structure identifiers
resulting from a substructure search, a chemical name search, a molecular
formula search, or by a combination of these searches. Structure classes
are an application of dynamics queries used to limit the scope of future
searches. For example, the system may maintain a structure class for
steroids. When a user performs a search, he or she can designate that the
result should be restricted to the members of the steroid class.
Accordingly, the user could simply query the database for all steroids that
have a particular substructure without drawing the entire steroid ring.
This results in two primary benefits: The first benefit is that queries
are simplified, i.e., there is no need to draw complex queries, and the
second benefit is that the screening phase need only be applied to those
compounds already known to be members of the structure class.
Dynamic queries and structure classes both exhibit a common benefit--the
overhead involved in structure searching is encountered only once (when the
dynamic query or structure class is defined) and additional overhead is
distributed evenly across subsequent updates to the chemical structure
database.
From the preceding description, it is evident that the invention has been
described in detail by reference to a particular embodiment adapted for use
in the field of chemistry. Although this invention offers many advantages
in this field, it may be used in other fields wherein structure data is
stored advantageously as well. Accordingly, this invention is not intended
to be limited by the details of the preferred embodiment described above,
but rather by the terms of the appended claims.
?
e as=scriptgen p
Ref Items Index-term
E1 1 AR=952
E2 1 AR=972
E3 0 *AS=SCRIPTGEN P
E4 236007 AU=A
E5 5 AU=A ADOLPH AMEN RA
E6 5 AU=A ADOLPH AMEN-RA
E7 1 AU=A AE BUTTED LYNG DORRIT A
E8 1 AU=A AE LYNG DORRIT A
E9 1 AU=A COSTA ANTHONY
E10 1 AU=A COURT ANDY
E11 1 AU=A HAANS P
E12 4 AU=A HEARN MICHAEL A
Enter P or PAGE for more
?
e pa=scriptgen
Ref Items Index-term
E1 9 PA=SCRIPTEL CORPORATION
E2 1 PA=SCRIPTEL HOLDING INC
E3 18 *PA=SCRIPTGEN
E4 18 PA=SCRIPTGEN PHARMACEUTICALS INC
E5 31 PA=SCRIPTO
E6 28 PA=SCRIPTO INC
E7 1 PA=SCRIPTO LLC
E8 1 PA=SCRIPTO TOKAI CORP
E9 1 PA=SCRIPTO TOKAI CORPORATION
E10 1 PA=SCRIPTO TOKAI INC
E11 1 PA=SCRIPTO-TOKAI CORPORATION
E12 1 PA=SCRIPTO-TOKAI INC
Enter P or PAGE for more
?
S E18;TYPE /PA,PN/ALL
S3 9 PA="SCRIPTPRO"
3/PA,PN/1 (Item 1 from file: 654)
DIALOG(R)File 654:(c) format only 2001 The Dialog Corp. All rts. reserv.
PATENT NO.: 6,161,721
ISSUED: December 19, 2000 (20001219)
ASSIGNEE(s): Scriptpro LLC, (A U.S. Company or Corporation), Mission, KS
(Kansas), US (United States of America)
[Assignee Code(s): 46711]
3/PA,PN/2 (Item 2 from file: 654)
DIALOG(R)File 654:(c) format only 2001 The Dialog Corp. All rts. reserv.
PATENT NO.: 6,155,485
ISSUED: December 05, 2000 (20001205)
ASSIGNEE(s): Scriptpro LLC, (A U.S. Company or Corporation), Mission, KS
(Kansas), US (United States of America)
[Assignee Code(s): 46711]
3/PA,PN/3 (Item 3 from file: 654)
DIALOG(R)File 654:(c) format only 2001 The Dialog Corp. All rts. reserv.
PATENT NO.: 6,085,938
ISSUED: July 11, 2000 (20000711)
ASSIGNEE(s): Scriptpro LLC, (A U.S. Company or Corporation), Mission, KS
(Kansas), US (United States of America)
[Assignee Code(s): 46711]
3/PA,PN/4 (Item 4 from file: 654)
DIALOG(R)File 654:(c) format only 2001 The Dialog Corp. All rts. reserv.
PATENT NO.: 5,897,024
ISSUED: April 27, 1999 (19990427)
ASSIGNEE(s): Scriptpro LLC, (A U.S. Company or Corporation), Mission, KS
(Kansas), US (United States of America)
[Assignee Code(s): 46711]
3/PA,PN/5 (Item 5 from file: 654)
DIALOG(R)File 654:(c) format only 2001 The Dialog Corp. All rts. reserv.
PATENT NO.: 5,873,488
ISSUED: February 23, 1999 (19990223)
ASSIGNEE(s): ScriptPro, LLC, (A U.S. Company or Corporation), Mission, KS
(Kansas), US (United States of America)
[Assignee Code(s): 46711]
3/PA,PN/6 (Item 6 from file: 654)
DIALOG(R)File 654:(c) format only 2001 The Dialog Corp. All rts. reserv.
PATENT NO.: 5,860,563
ISSUED: January 19, 1999 (19990119)
ASSIGNEE(s): Scriptpro, LLC, (A U.S. Company or Corporation), Mission, KS
(Kansas), US (United States of America)
[Assignee Code(s): 46711]
3/PA,PN/7 (Item 7 from file: 654)
DIALOG(R)File 654:(c) format only 2001 The Dialog Corp. All rts. reserv.
PATENT NO.: 5,798,020
ISSUED: August 25, 1998 (19980825)
ASSIGNEE(s): Scriptpro, LLC, (A U.S. Company or Corporation), Mission, KS
(Kansas), US (United States of America)
[Assignee Code(s): 46711]
3/PA,PN/8 (Item 8 from file: 654)
DIALOG(R)File 654:(c) format only 2001 The Dialog Corp. All rts. reserv.
PATENT NO.: 5,762,235
ISSUED: June 09, 1998 (19980609)
ASSIGNEE(s): ScriptPro, L L C , (A U.S. Company or Corporation), Shawnee
Mission, KS (Kansas), US (United States of America)
[Assignee Code(s): 46711]
3/PA,PN/9 (Item 9 from file: 654)
DIALOG(R)File 654:(c) format only 2001 The Dialog Corp. All rts. reserv.
PATENT NO.: 5,713,487
ISSUED: February 03, 1998 (19980203)
ASSIGNEE(s): ScriptPro L L C , (A U.S. Company or Corporation), Shawnee
Mission, KS (Kansas), US (United States of America)
?
S E4/PA,PN/ALL
>>>Possible typing error near /
?
S E4; T /PA,PN/ALL
S4 18 PA="SCRIPTGEN PHARMACEUTICALS INC"
4/PA,PN/1 (Item 1 from file: 654)
DIALOG(R)File 654:(c) format only 2001 The Dialog Corp. All rts. reserv.
PATENT NO.: 6,239,291
ISSUED: May 29, 2001 (20010529)
ASSIGNEE(s): Scriptgen Pharmaceuticals, Inc , (A U.S. Company or
Corporation), Waltham, MA (Massachusetts), US (United States
of America)
Trustees of Boston University, (A U.S. Company or Corporation)
, Boston, MA (Massachusetts), US (United States of America)
[Assignee Code(s): 1308; 40527]
4/PA,PN/2 (Item 2 from file: 654)
DIALOG(R)File 654:(c) format only 2001 The Dialog Corp. All rts. reserv.
PATENT NO.: 6,228,882
ISSUED: May 08, 2001 (20010508)
ASSIGNEE(s): Anadys Pharmaceuticals, Inc , (A U.S. Company or Corporation),
Boston, MA (Massachusetts), US (United States of America)
Scriptgen Pharmaceuticals, Inc , (A U.S. Company or
Corporation), Waltham, MA (Massachusetts), US (United States
of America)
[Assignee Code(s): 40527; 56074]
4/PA,PN/3 (Item 3 from file: 654)
DIALOG(R)File 654:(c) format only 2001 The Dialog Corp. All rts. reserv.
PATENT NO.: 6,194,399
ISSUED: February 27, 2001 (20010227)
ASSIGNEE(s): Scriptgen Pharmaceuticals, Inc , (A U.S. Company or
Corporation), Waltham, MA (Massachusetts), US (United States
of America)
[Assignee Code(s): 40527]
4/PA,PN/4 (Item 4 from file: 654)
DIALOG(R)File 654:(c) format only 2001 The Dialog Corp. All rts. reserv.
PATENT NO.: 6,184,393
ISSUED: February 06, 2001 (20010206)
ASSIGNEE(s): Scriptgen Pharmaceuticals, Inc , (A U.S. Company or
Corporation), Waltham, MA (Massachusetts), US (United States
of America)
Trustees of Boston University, (A U.S. Company or Corporation)
, Boston, MA (Massachusetts), US (United States of America)
[Assignee Code(s): 1308; 40527]
4/PA,PN/5 (Item 5 from file: 654)
DIALOG(R)File 654:(c) format only 2001 The Dialog Corp. All rts. reserv.
PATENT NO.: 6,165,998
ISSUED: December 26, 2000 (20001226)
ASSIGNEE(s): Scriptgen Pharmaceuticals, Inc , (A U.S. Company or
Corporation), Waltham, MA (Massachusetts), US (United States
of America)
[Assignee Code(s): 40527]
4/PA,PN/6 (Item 6 from file: 654)
DIALOG(R)File 654:(c) format only 2001 The Dialog Corp. All rts. reserv.
PATENT NO.: 6,140,361
ISSUED: October 31, 2000 (20001031)
ASSIGNEE(s): Scriptgen Pharmaceuticals, Inc , (A U.S. Company or
Corporation), Waltham, MA (Massachusetts), US (United States
of America)
[Assignee Code(s): 40527]
4/PA,PN/7 (Item 7 from file: 654)
DIALOG(R)File 654:(c) format only 2001 The Dialog Corp. All rts. reserv.
PATENT NO.: 6,127,551
ISSUED: October 03, 2000 (20001003)
ASSIGNEE(s): Scriptgen Pharmaceuticals, Inc , (A U.S. Company or
Corporation), Waltham, MA (Massachusetts), US (United States
of America)
Trustees of Boston University, (A U.S. Company or Corporation)
, Boston, MA (Massachusetts), US (United States of America)
[Assignee Code(s): 1308; 40527]
4/PA,PN/8 (Item 8 from file: 654)
DIALOG(R)File 654:(c) format only 2001 The Dialog Corp. All rts. reserv.
PATENT NO.: 6,090,953
ISSUED: July 18, 2000 (20000718)
ASSIGNEE(s): Scriptgen Pharmaceuticals, Inc , (A U.S. Company or
Corporation), Boston, MA (Massachusetts), US (United States of
America)
Trustees of Boston University, (A U.S. Company or Corporation)
, Waltham, MA (Massachusetts), US (United States of America)
[Assignee Code(s): 1308; 40527]
4/PA,PN/9 (Item 9 from file: 654)
DIALOG(R)File 654:(c) format only 2001 The Dialog Corp. All rts. reserv.
PATENT NO.: 6,051,373
ISSUED: April 18, 2000 (20000418)
ASSIGNEE(s): Scriptgen Pharmaceuticals, Inc , (A U.S. Company or
Corporation), Waltham, MA (Massachusetts), US (United States
of America)
University of Massachusetts Medical Center, (A U.S. Company or
Corporation), Worcester, MA (Massachusetts), US (United States
of America)
[Assignee Code(s): 22237; 40527]
4/PA,PN/10 (Item 10 from file: 654)
DIALOG(R)File 654:(c) format only 2001 The Dialog Corp. All rts. reserv.
PATENT NO.: 6,022,983
ISSUED: February 08, 2000 (20000208)
ASSIGNEE(s): Scriptgen Pharmaceuticals, Inc , (A U.S. Company or
Corporation), Waltham, MA (Massachusetts), US (United States
of America)
Trustees of Boston University, (A U.S. Company or Corporation)
, Boston, MA (Massachusetts), US (United States of America)
[Assignee Code(s): 1308; 40527]
4/PA,PN/11 (Item 11 from file: 654)
DIALOG(R)File 654:(c) format only 2001 The Dialog Corp. All rts. reserv.
PATENT NO.: 6,020,488
ISSUED: February 01, 2000 (20000201)
ASSIGNEE(s): Scriptgen Pharmaceuticals Inc , (A U.S. Company or
Corporation), Waltham, MA (Massachusetts), US (United States
of America)
[Assignee Code(s): 40527]
4/PA,PN/12 (Item 12 from file: 654)
DIALOG(R)File 654:(c) format only 2001 The Dialog Corp. All rts. reserv.
PATENT NO.: 6,004,779
ISSUED: December 21, 1999 (19991221)
ASSIGNEE(s): Scriptgen Pharmaceuticals, Inc , (A U.S. Company or
Corporation), Waltham, MA (Massachusetts), US (United States
of America)
[Assignee Code(s): 40527]
4/PA,PN/13 (Item 13 from file: 654)
DIALOG(R)File 654:(c) format only 2001 The Dialog Corp. All rts. reserv.
PATENT NO.: 5,986,111
ISSUED: November 16, 1999 (19991116)
ASSIGNEE(s): Scriptgen Pharmaceuticals, Inc , (A U.S. Company or
Corporation), Waltham, MA (Massachusetts), US (United States
of America)
Trustees of Boston University, (A U.S. Company or Corporation)
, Boston, MA (Massachusetts), US (United States of America)
[Assignee Code(s): 1308; 40527]
4/PA,PN/14 (Item 14 from file: 654)
DIALOG(R)File 654:(c) format only 2001 The Dialog Corp. All rts. reserv.
PATENT NO.: 5,942,547
ISSUED: August 24, 1999 (19990824)
ASSIGNEE(s): Scriptgen Pharmaceuticals, Inc , (A U.S. Company or
Corporation), Waltham, MA (Massachusetts), US (United States
of America)
[Assignee Code(s): 40527]
4/PA,PN/15 (Item 15 from file: 654)
DIALOG(R)File 654:(c) format only 2001 The Dialog Corp. All rts. reserv.
PATENT NO.: 5,760,063
ISSUED: June 02, 1998 (19980602)
ASSIGNEE(s): Scriptgen Pharmaceuticals, Inc , (A U.S. Company or
Corporation), Medford, MA (Massachusetts), US (United States
of America)
[Assignee Code(s): 40527]
4/PA,PN/16 (Item 16 from file: 654)
DIALOG(R)File 654:(c) format only 2001 The Dialog Corp. All rts. reserv.
PATENT NO.: 5,679,582
ISSUED: October 21, 1997 (19971021)
ASSIGNEE(s): Scriptgen Pharmaceuticals, Inc , (A U.S. Company or
Corporation), Medford, MA (Massachusetts), US (United States
of America)
[Assignee Code(s): 40527]
4/PA,PN/17 (Item 17 from file: 654)
DIALOG(R)File 654:(c) format only 2001 The Dialog Corp. All rts. reserv.
PATENT NO.: 5,668,165
ISSUED: September 16, 1997 (19970916)
ASSIGNEE(s): Scriptgen Pharmaceuticals, Inc , (A U.S. Company or
Corporation), Medford, MA (Massachusetts), US (United States
of America)
[Assignee Code(s): 40527]
4/PA,PN/18 (Item 18 from file: 654)
DIALOG(R)File 654:(c) format only 2001 The Dialog Corp. All rts. reserv.
PATENT NO.: 5,585,277
ISSUED: December 17, 1996 (19961217)
ASSIGNEE(s): Scriptgen Pharmaceuticals, Inc, (A U.S. Company or
Corporation), Medford, MA (Massachusetts), US (United States
of America)
[Assignee Code(s): 40527]
?
e pa=cubist p
Ref Items Index-term
E1 1 PA=CUBIFORM DESIGN AND DEV CO INC CN
E2 14 PA=CUBIST
E3 0 *PA=CUBIST P
E4 14 PA=CUBIST PHARMACEUTICALS INC
E5 3 PA=CUBIT
E6 2 PA=CUBIT CORP
E7 2 PA=CUBIT CORPORATION
E8 1 PA=CUBIT LIMITED
E9 1 PA=CUBIT LTD GB
E10 10 PA=CUBITAL
E11 10 PA=CUBITAL LTD
E12 10 PA=CUBITAL LTD IL
Enter P or PAGE for more
?
S E4; T/PN,PA/ALL
S5 14 PA="CUBIST PHARMACEUTICALS INC"
5/PN,PA/1 (Item 1 from file: 654)
DIALOG(R)File 654:(c) format only 2001 The Dialog Corp. All rts. reserv.
PATENT NO.: 6,221,640
ISSUED: April 24, 2001 (20010424)
ASSIGNEE(s): Cubist Pharmaceuticals, Inc , (A U.S. Company or Corporation),
Cambridge, MA (Massachusetts), US (United States of America)
[Assignee Code(s): 41839]
5/PN,PA/2 (Item 2 from file: 654)
DIALOG(R)File 654:(c) format only 2001 The Dialog Corp. All rts. reserv.
PATENT NO.: 6,174,713
ISSUED: January 16, 2001 (20010116)
ASSIGNEE(s): Cubist Pharmaceuticals, Inc , (A U.S. Company or Corporation),
Cambridge, MA (Massachusetts), US (United States of America)
[Assignee Code(s): 41839]
5/PN,PA/3 (Item 3 from file: 654)
DIALOG(R)File 654:(c) format only 2001 The Dialog Corp. All rts. reserv.
PATENT NO.: 6,153,645
ISSUED: November 28, 2000 (20001128)
ASSIGNEE(s): Cubist Pharmaceuticals, Inc , (A U.S. Company or Corporation),
Cambridge, MA (Massachusetts), US (United States of America)
[Assignee Code(s): 41839]
5/PN,PA/4 (Item 4 from file: 654)
DIALOG(R)File 654:(c) format only 2001 The Dialog Corp. All rts. reserv.
PATENT NO.: 5,912,140
ISSUED: June 15, 1999 (19990615)
ASSIGNEE(s): Cubist Pharmaceuticals, Inc , (A U.S. Company or Corporation),
Cambridge, MA (Massachusetts), US (United States of America)
[Assignee Code(s): 41839]
5/PN,PA/5 (Item 5 from file: 654)
DIALOG(R)File 654:(c) format only 2001 The Dialog Corp. All rts. reserv.
PATENT NO.: 5,885,815
ISSUED: March 23, 1999 (19990323)
ASSIGNEE(s): Cubist Pharmaceuticals, Inc , (A U.S. Company or Corporation),
Cambridge, MA (Massachusetts), US (United States of America)
[Assignee Code(s): 41839]
5/PN,PA/6 (Item 6 from file: 654)
DIALOG(R)File 654:(c) format only 2001 The Dialog Corp. All rts. reserv.
PATENT NO.: 5,871,987
ISSUED: February 16, 1999 (19990216)
ASSIGNEE(s): Cubist Pharmaceuticals, Inc , (A U.S. Company or Corporation),
Cambridge, MA (Massachusetts), US (United States of America)
[Assignee Code(s): 41839]
5/PN,PA/7 (Item 7 from file: 654)
DIALOG(R)File 654:(c) format only 2001 The Dialog Corp. All rts. reserv.
PATENT NO.: 5,824,657
ISSUED: October 20, 1998 (19981020)
ASSIGNEE(s): Cubist Pharmaceuticals, Inc , (A U.S. Company or Corporation),
Cambridge, MA (Massachusetts), US (United States of America)
[Assignee Code(s): 41839]
5/PN,PA/8 (Item 8 from file: 654)
DIALOG(R)File 654:(c) format only 2001 The Dialog Corp. All rts. reserv.
PATENT NO.: 5,801,013
ISSUED: September 01, 1998 (19980901)
ASSIGNEE(s): Cubist Pharmaceuticals, Inc , (A U.S. Company or Corporation),
Cambridge, MA (Massachusetts), US (United States of America)
[Assignee Code(s): 41839]
5/PN,PA/9 (Item 9 from file: 654)
DIALOG(R)File 654:(c) format only 2001 The Dialog Corp. All rts. reserv.
PATENT NO.: 5,798,240
ISSUED: August 25, 1998 (19980825)
ASSIGNEE(s): Cubist Pharmaceuticals, Inc , (A U.S. Company or Corporation),
Cambridge, MA (Massachusetts), US (United States of America)
[Assignee Code(s): 41839]
5/PN,PA/10 (Item 10 from file: 654)
DIALOG(R)File 654:(c) format only 2001 The Dialog Corp. All rts. reserv.
PATENT NO.: 5,759,833
ISSUED: June 02, 1998 (19980602)
ASSIGNEE(s): Cancer Institute, Japanese Foundation for Cancer Research, (A
Non-U.S. Company or Corporation), Tokyo, JP (Japan)
Cubist Pharmaceuticals, Inc , (A U.S. Company or Corporation),
Cambridge, MA (Massachusetts), US (United States of America)
[Assignee Code(s): 12629; 33923; 41839]
5/PN,PA/11 (Item 11 from file: 654)
DIALOG(R)File 654:(c) format only 2001 The Dialog Corp. All rts. reserv.
PATENT NO.: 5,756,327
ISSUED: May 26, 1998 (19980526)
ASSIGNEE(s): Cubist Pharmaceuticals, Inc , (A U.S. Company or Corporation),
Cambridge, MA (Massachusetts), US (United States of America)
[Assignee Code(s): 41839]
5/PN,PA/12 (Item 12 from file: 654)
DIALOG(R)File 654:(c) format only 2001 The Dialog Corp. All rts. reserv.
PATENT NO.: 5,726,195
ISSUED: March 10, 1998 (19980310)
ASSIGNEE(s): Cubist Pharmaceuticals, Inc , (A U.S. Company or Corporation),
Cambridge, MA (Massachusetts), US (United States of America)
[Assignee Code(s): 41839]
5/PN,PA/13 (Item 13 from file: 654)
DIALOG(R)File 654:(c) format only 2001 The Dialog Corp. All rts. reserv.
PATENT NO.: 5,656,470
ISSUED: August 12, 1997 (19970812)
ASSIGNEE(s): Cubist Pharmaceuticals, Inc, (A U.S. Company or Corporation),
Cambridge, MA (Massachusetts), US (United States of America)
[Assignee Code(s): 41839]
5/PN,PA/14 (Item 14 from file: 654)
DIALOG(R)File 654:(c) format only 2001 The Dialog Corp. All rts. reserv.
PATENT NO.: 5,629,188
ISSUED: May 13, 1997 (19970513)
ASSIGNEE(s): Cancer Institute, Japanese Foundation for Cancer Research, (A
Non-U.S. Company or Corporation), Tokyo, JP (Japan)
Cubist Pharmaceuticals, Inc, (A U.S. Company or Corporation),
Cambridge, MA (Massachusetts), US (United States of America)
Massachusetts Institute of Technology, (A U.S. Company or
Corporation), Cambridge, MA (Massachusetts), US (United States
of America)
[Assignee Code(s): 12629; 33923; 41839; 52912]
?
TYPE /PN,TI,PA/ALL
5/PN,TI,PA/1 (Item 1 from file: 654)
DIALOG(R)File 654:(c) format only 2001 The Dialog Corp. All rts. reserv.
ENTEROCOCCAL AMINOACYL-TRNA SYNTHETASE PROTEINS, NUCLEIC ACIDS AND STRAINS
COMPRISING SAME
PATENT NO.: 6,221,640
ISSUED: April 24, 2001 (20010424)
ASSIGNEE(s): Cubist Pharmaceuticals, Inc , (A U.S. Company or Corporation),
Cambridge, MA (Massachusetts), US (United States of America)
[Assignee Code(s): 41839]
5/PN,TI,PA/2 (Item 2 from file: 654)
DIALOG(R)File 654:(c) format only 2001 The Dialog Corp. All rts. reserv.
CANDIDA CYTOPLASMIC TRYPTOPHANYL-TRNA SYNTHETASE PROTEINS, NUCLEIC ACIDS
AND STRAINS COMPRISING SAME
PATENT NO.: 6,174,713
ISSUED: January 16, 2001 (20010116)
ASSIGNEE(s): Cubist Pharmaceuticals, Inc , (A U.S. Company or Corporation),
Cambridge, MA (Massachusetts), US (United States of America)
[Assignee Code(s): 41839]
5/PN,TI,PA/3 (Item 3 from file: 654)
DIALOG(R)File 654:(c) format only 2001 The Dialog Corp. All rts. reserv.
HETEROCYCLES AS ANTIMICROBIAL AGENTS
PATENT NO.: 6,153,645
ISSUED: November 28, 2000 (20001128)
ASSIGNEE(s): Cubist Pharmaceuticals, Inc , (A U.S. Company or Corporation),
Cambridge, MA (Massachusetts), US (United States of America)
[Assignee Code(s): 41839]
5/PN,TI,PA/4 (Item 4 from file: 654)
DIALOG(R)File 654:(c) format only 2001 The Dialog Corp. All rts. reserv.
RECOMBINANT PNEUMOCYSTIS CARINII AMINOACYL TRNA SYNTHETASE GENES, TESTER
STRAINS AND ASSAYS
[ Isolated nucleic acid which codes a functional portion of a lysyl-tRNA
synthetase; having catalytic acitivity and binding function; drug screening
for AIDS therapy]
PATENT NO.: 5,912,140
ISSUED: June 15, 1999 (19990615)
ASSIGNEE(s): Cubist Pharmaceuticals, Inc , (A U.S. Company or Corporation),
Cambridge, MA (Massachusetts), US (United States of America)
[Assignee Code(s): 41839]
5/PN,TI,PA/5 (Item 5 from file: 654)
DIALOG(R)File 654:(c) format only 2001 The Dialog Corp. All rts. reserv.
CANDIDA ISOLEUCYL-TRNA SYNTHETASE PROTEINS, NUCLEIC ACIDS AND STRAINS
COMPRISING SAME
[Nucleic acid encoding Candida isoleucyl-tRNA synthetase]
PATENT NO.: 5,885,815
ISSUED: March 23, 1999 (19990323)
ASSIGNEE(s): Cubist Pharmaceuticals, Inc , (A U.S. Company or Corporation),
Cambridge, MA (Massachusetts), US (United States of America)
[Assignee Code(s): 41839]
5/PN,TI,PA/6 (Item 6 from file: 654)
DIALOG(R)File 654:(c) format only 2001 The Dialog Corp. All rts. reserv.
CANDIDA TYROSYL-TRNA SYNTHETASE PROTEINS, NUCLEIC ACIDS AND STRAINS
COMPRISING SAME
PATENT NO.: 5,871,987
ISSUED: February 16, 1999 (19990216)
ASSIGNEE(s): Cubist Pharmaceuticals, Inc , (A U.S. Company or Corporation),
Cambridge, MA (Massachusetts), US (United States of America)
[Assignee Code(s): 41839]
5/PN,TI,PA/7 (Item 7 from file: 654)
DIALOG(R)File 654:(c) format only 2001 The Dialog Corp. All rts. reserv.
AMINOACYL SULFAMIDES FOR THE TREATMENT OF HYPERPROLIFERATIVE DISORDERS
[Anticarcinogenic agents, skin disorders and psoriasis]
PATENT NO.: 5,824,657
ISSUED: October 20, 1998 (19981020)
ASSIGNEE(s): Cubist Pharmaceuticals, Inc , (A U.S. Company or Corporation),
Cambridge, MA (Massachusetts), US (United States of America)
[Assignee Code(s): 41839]
5/PN,TI,PA/8 (Item 8 from file: 654)
DIALOG(R)File 654:(c) format only 2001 The Dialog Corp. All rts. reserv.
HELICOBACTER AMINOACYL-TRNA SYNTHETASE PROTEINS, NUCLEIC ACIDS AND STRAINS
COMPRISING SAME
PATENT NO.: 5,801,013
ISSUED: September 01, 1998 (19980901)
ASSIGNEE(s): Cubist Pharmaceuticals, Inc , (A U.S. Company or Corporation),
Cambridge, MA (Massachusetts), US (United States of America)
[Assignee Code(s): 41839]
5/PN,TI,PA/9 (Item 9 from file: 654)
DIALOG(R)File 654:(c) format only 2001 The Dialog Corp. All rts. reserv.
RECOMBINANT MYCOBACTERIAL METHIONYL-TRNA SYNTHETASE GENES AND METHODS OF
USE THEREFORE
[Isolated nucleic acid encoding aminoacyl-transfer RNA synthetase]
PATENT NO.: 5,798,240
ISSUED: August 25, 1998 (19980825)
ASSIGNEE(s): Cubist Pharmaceuticals, Inc , (A U.S. Company or Corporation),
Cambridge, MA (Massachusetts), US (United States of America)
[Assignee Code(s): 41839]
5/PN,TI,PA/10 (Item 10 from file: 654)
DIALOG(R)File 654:(c) format only 2001 The Dialog Corp. All rts. reserv.
HUMAN ISOLEUCYL-TRNA SYNTHETASE PROTEINS, NUCLEIC ACIDS AND TESTER STRAINS
COMPRISING SAME
PATENT NO.: 5,759,833
ISSUED: June 02, 1998 (19980602)
ASSIGNEE(s): Cancer Institute, Japanese Foundation for Cancer Research, (A
Non-U.S. Company or Corporation), Tokyo, JP (Japan)
Cubist Pharmaceuticals, Inc , (A U.S. Company or Corporation),
Cambridge, MA (Massachusetts), US (United States of America)
[Assignee Code(s): 12629; 33923; 41839]
5/PN,TI,PA/11 (Item 11 from file: 654)
DIALOG(R)File 654:(c) format only 2001 The Dialog Corp. All rts. reserv.
RECOMBINANT MYCOBACTERIAL ISOLEUCYL-TRNA SYNTHETASE GENES, TESTER STRAINS
AND ASSAYS
[ Isolated nucleic acid encoding isoleucyl-transferRNA synthetase of genus
Mycobacterium]
PATENT NO.: 5,756,327
ISSUED: May 26, 1998 (19980526)
ASSIGNEE(s): Cubist Pharmaceuticals, Inc , (A U.S. Company or Corporation),
Cambridge, MA (Massachusetts), US (United States of America)
[Assignee Code(s): 41839]
5/PN,TI,PA/12 (Item 12 from file: 654)
DIALOG(R)File 654:(c) format only 2001 The Dialog Corp. All rts. reserv.
AMINOACYL ADENYLATE MIMICS AS NOVEL ANTIMICROBIAL AND ANTIPARASITIC AGENTS
PATENT NO.: 5,726,195
ISSUED: March 10, 1998 (19980310)
ASSIGNEE(s): Cubist Pharmaceuticals, Inc , (A U.S. Company or Corporation),
Cambridge, MA (Massachusetts), US (United States of America)
[Assignee Code(s): 41839]
5/PN,TI,PA/13 (Item 13 from file: 654)
DIALOG(R)File 654:(c) format only 2001 The Dialog Corp. All rts. reserv.
RECOMBINANT MYCOBACTERIAL SERYL-TRNA SYNTHETASE GENES, TESTER STRAINS AND
ASSAYS
PATENT NO.: 5,656,470
ISSUED: August 12, 1997 (19970812)
ASSIGNEE(s): Cubist Pharmaceuticals, Inc, (A U.S. Company or Corporation),
Cambridge, MA (Massachusetts), US (United States of America)
[Assignee Code(s): 41839]
5/PN,TI,PA/14 (Item 14 from file: 654)
DIALOG(R)File 654:(c) format only 2001 The Dialog Corp. All rts. reserv.
HUMAN ALANYL-TRNA SYNTHETASE PROTEINS, NUCLEIC ACIDS AND TESTER STRAINS
COMPRISING SAME
[ Methods for making human alanyl-tRNA synthetase, e.g., maintaining host
cells comprising recombinant gene encoding synthetase under conditions in
which gene is expressed, isolating synthetase; related vectors, plasmid]
PATENT NO.: 5,629,188
ISSUED: May 13, 1997 (19970513)
ASSIGNEE(s): Cancer Institute, Japanese Foundation for Cancer Research, (A
Non-U.S. Company or Corporation), Tokyo, JP (Japan)
Cubist Pharmaceuticals, Inc, (A U.S. Company or Corporation),
Cambridge, MA (Massachusetts), US (United States of America)
Massachusetts Institute of Technology, (A U.S. Company or
Corporation), Cambridge, MA (Massachusetts), US (United States
of America)
[Assignee Code(s): 12629; 33923; 41839; 52912]
?
B PATENTS
>>> 123 is unauthorized
>>> 340 is unauthorized
>>> 344 is unauthorized
>>> 345 is unauthorized
>>> 351 is unauthorized
>>> 352 is unauthorized
>>> 353 is unauthorized
>>> 447 is unauthorized
>>> 670 is unauthorized
>>>9 of the specified files are not available
19sep01 18:16:28 User726734 Session D434.3
$0.32 0.212 DialUnits File653
$0.00 1 Type(s) in Format 2
$0.00 1 Type(s) in Format 9 (UDF)
$0.00 2 Types
$0.32 Estimated cost File653
$0.18 0.120 DialUnits File652
$0.18 Estimated cost File652
$4.98 3.321 DialUnits File654
$0.00 2 Type(s) in Format 2
$0.00 55 Type(s) in Format 2 (UDF)
$0.00 2 Type(s) in Format 9 (UDF)
$0.00 59 Types
$4.98 Estimated cost File654
OneSearch, 3 files, 3.654 DialUnits FileOS
$0.60 INTERNET
$6.08 Estimated cost this search
$6.87 Estimated total session cost 4.149 DialUnits
SYSTEM:OS - DIALOG OneSearch
File 342:Derwent Patents Citation Indx 1978-01/200147
(c) 2001 Derwent Info Ltd
*File 342: Price changes as of 1/1/01. Please see HELP RATES 342.
File 347:JAPIO OCT 1976-2001/May(UPDATED 010905)
(c) 2001 JPO & JAPIO
*File 347: JAPIO data problems with year 2000 records are now fixed.
Alerts have been run. See HELP NEWS 347 for details.
File 348:EUROPEAN PATENTS 1978-2001/Sep W02
(c) 2001 European Patent Office
File 349:PCT Fulltext 1983-2001/UB=20010906, UT=20010830
(c) 2001 WIPO/MicroPat
File 371:French Patents 1961-2001/BOPI 200136
(c) 2001 INPI. All rts. reserv.
File 652:US Patents Fulltext 1971-1979
(c) format only 2001 The Dialog Corp.
*File 652: Reassignment data current through June 6, 2001 recordings.
Due to processing problems, the SORT command is not working.
File 653:US Patents Fulltext 1980-1989
(c) format only 2001 The Dialog Corp.
*File 653: Reassignment data current through June 6, 2001 recordings.
Due to processing problems, the SORT command is not working.
File 654:US PAT.FULL. 1990-2001/Sep 18
(c) format only 2001 The Dialog Corp.
*File 654: Reassignment data current through June 6, 2001 recordings
Set Items Description
--- ----- -----------
?
B 342
19sep01 18:16:47 User726734 Session D434.4
$0.03 0.022 DialUnits File342
$0.03 Estimated cost File342
$0.03 0.022 DialUnits File347
$0.03 Estimated cost File347
$0.03 0.022 DialUnits File348
$0.03 Estimated cost File348
$0.03 0.022 DialUnits File349
$0.03 Estimated cost File349
$0.03 0.022 DialUnits File371
$0.03 Estimated cost File371
$0.03 0.022 DialUnits File652
$0.03 Estimated cost File652
$0.03 0.022 DialUnits File653
$0.03 Estimated cost File653
$0.03 0.022 DialUnits File654
$0.03 Estimated cost File654
OneSearch, 8 files, 0.173 DialUnits FileOS
$0.02 INTERNET
$0.26 Estimated cost this search
$7.13 Estimated total session cost 4.321 DialUnits
File 342:Derwent Patents Citation Indx 1978-01/200147
(c) 2001 Derwent Info Ltd
*File 342: Price changes as of 1/1/01. Please see HELP RATES 342.
Set Items Description
--- ----- -----------
?
?
S PN=US 4642762 OR PN=US 5577239 OR PN=US 5950192
1 PN=US 4642762
1 PN=US 5577239
1 PN=US 5950192
S2 2 PN=US 4642762 OR PN=US 5577239 OR PN=US 5950192
?
TYPE 2/2/ALL
2/2/1
DIALOG(R)File 342:Derwent Patents Citation Indx
(c) 2001 Derwent Info Ltd. All rts. reserv.
02412679 WPI Acc No: 96-151525/15
Chemical structure storage method using relational database - storing
matrix of chemical structure including atoms and bonds in relational
database table and generating and storing search keys for each atom in
chemical structure
Patent Assignee: (PSII-) PSI INT INC
Author (Inventor): MOORE J; BRAZIL J; HOOVER J R
Patent Family:
Patent No Kind Date Examiner Field of Search
WO 9606391 A2 960229 (BASIC) None
AU 9533202 A 960314
EP 777882 A1 970611
EP 777882 A4 971229
JP 3193383 B2 010730
JP 10507285 W 980714
US 5577239 A 961119 364/496; 364/497; 364/499; 364/DIG.1; 364/DIG.2;
395/600
US 5950192 A 990907 395/496; 395/497; 395/499; 395/600; 395/603;
702/19; 702/20; 702/22; 702/27; 707/1;
707/100; 707/102; 707/104; 707/19; 707/20;
707/22; 707/27; 707/3
WO 9606391 A3 960509 364/496; 364/497; 364/499; 395/600
Derwent Week (Basic): 9615
Priority Data: US 288503 (940810)
Applications: US 288503 (940810); AU 9533202 (950810); EP 95929457 (950810
); WO 95US10171 (950810); JP 96508133 (950810); US 883165 (970626)
Designated States
(National): AM; AT; AU; BB; BG; BR; BY; CA; CH; CN; CZ; DE; DK; ES; FI;
GB; GE; HU; JP; KE; KG; KP; KR; KZ; LK; LT; LU; LV; MD; MG; MN; MW; MX
; NO; NZ; PL; PT; RO; RU; SD; SE; SI; SK; TJ; TT; UA; US; UZ; VN
(Regional): AT; BE; CH; DE; DK; ES; FR; GB; GR; IE; IT; KE; LI; LU; MC;
MW; NL; OA; PT; SD; SE; SZ; UG
Derwent Class: T01
Int Pat Class: G06F-017/30
Number of Patents: 009
Number of Countries: 060
Number of Cited Patents: 036
Number of Cited Literature References: 039
Number of Citing Patents: 006
CITED PATENTS
Family Member Cited Patent Cat WPI Acc No Assignee/Inventor
By Examiner:
EP 777882 A EP 90895 A A 83-790076/42 (NIIN-) JAPAN INFORM CENT
/ARAKI K
EP 777882 A EP 213483 A A 87-066235/10 (FUJF ) FUJI PHOTO FILM
CO LTD/FUJITA S
EP 777882 A US 4642762 A A 87-056518/08 (AMCH-) AMER CHEMICAL SOC
/FISANICK W
US 5577239 A US 4642762 A 87-056518/08 (AMCH-) AMER CHEMICAL SOC
/FISANICK W
US 5577239 A US 4811217 A 86-259665/40 (NIAS-) JAPAN ASSOC INT
CHE/TOKIZANE S; CHIHARA H
US 5577239 A US 4855931 A 90-014420/02 (UYYA ) UNIV YALE/
SAUNDERS M
US 5577239 A US 5025388 A 91-200832/27 (CRAM/) CRAMER R D/CRAMER
R D; WOLD S V
US 5577239 A US 5056035 A 91-317888/43 (FUJF ) FUJI PHOTO FILM
CO LTD/FUJITA S
US 5577239 A US 5249137 A 91-283380/39 (XERO ) XEROX CORP/KAPLAN
S; MALLGREEN W R; FACCI J S; DONALDSON J M
US 5577239 A US 5367058 A 95-005895/01 (BECT ) BECTON DICKINSON
CO/MIZE P D; PITNER J B; LINN
US 5577239 A US 5379234 A 91-283380/39 (XERO ) XEROX CORP/KAPLAN
S; MALLGREEN W R; FACCI J S; DONALDSON J M
US 5577239 A US 5386507 A 95-081830/11 (KAHN/) KAHN S D; (TEIG/)
TEIG S L/TEIG S L; KAHN S D
US 5577239 A US 5418944 A 92-260461/32 (IBMC ) INT BUSINESS
MACHINES CORP; (IBMC ) IBM SEMEA SRL/FABROCINI F;
DI PACE L
US 5577239 A US 5463564 A 95-392190/50 (THRE-) 3-DIMENSIONAL
PHARM INC/BONE R F; AGRAFIOTIS D K; SALEMME F R;
SOLL R M
US 5950192 A EP 90895 A2 83-790076/42 (NIIN-) JAPAN INFORM CENT
/ARAKI K
US 5950192 A EP 213483 A2 87-066235/10 (FUJF ) FUJI PHOTO FILM
CO LTD/FUJITA S
US 5950192 A US 4642762 A 87-056518/08 (AMCH-) AMER CHEMICAL SOC
/FISANICK W
US 5950192 A US 4811217 A 86-259665/40 (NIAS-) JAPAN ASSOC INT
CHE/TOKIZANE S; CHIHARA H
US 5950192 A US 4855931 A 90-014420/02 (UYYA ) UNIV YALE/
SAUNDERS M
US 5950192 A US 5025388 A 91-200832/27 (CRAM/) CRAMER R D/CRAMER
R D; WOLD S V
US 5950192 A US 5056035 A 91-317888/43 (FUJF ) FUJI PHOTO FILM
CO LTD/FUJITA S
US 5950192 A US 5259137 A 93-367341/46 (BLAS-) BLASER
JAGDWAFFENFABRIK HORST/BLENK G; ZEH M
US 5950192 A US 5367058 A 95-005895/01 (BECT ) BECTON DICKINSON
CO/MIZE P D; PITNER J B; LINN
US 5950192 A US 5379234 A 91-283380/39 (XERO ) XEROX CORP/KAPLAN
S; MALLGREEN W R; FACCI J S; DONALDSON J M
US 5950192 A US 5386507 A 95-081830/11 (KAHN/) KAHN S D; (TEIG/)
TEIG S L/TEIG S L; KAHN S D
US 5950192 A US 5418944 A 92-260461/32 (IBMC ) INT BUSINESS
MACHINES CORP; (IBMC ) IBM SEMEA SRL/FABROCINI F;
DI PACE L
US 5950192 A US 5463564 A 95-392190/50 (THRE-) 3-DIMENSIONAL
PHARM INC/BONE R F; AGRAFIOTIS D K; SALEMME F R;
SOLL R M
US 5950192 A US 5577239 A 96-151525/15 (PSII-) PSI INT INC/MOORE
J; BRAZIL J; HOOVER J R
WO 9606391 A US 4811217 A X 86-259665/40 (NIAS-) JAPAN ASSOC INT
CHE/TOKIZANE S; CHIHARA H
WO 9606391 A US 4855931 A A 90-014420/02 (UYYA ) UNIV YALE/
SAUNDERS M
WO 9606391 A US 5025388 A A 91-200832/27 (CRAM/) CRAMER R D/CRAMER
R D; WOLD S V
WO 9606391 A US 5056035 A A 91-317888/43 (FUJF ) FUJI PHOTO FILM
CO LTD/FUJITA S
WO 9606391 A3 US 4811217 A X 86-259665/40 (NIAS-) JAPAN ASSOC INT
CHE/TOKIZANE S; CHIHARA H
WO 9606391 A3 US 4855931 A A 90-014420/02 (UYYA ) UNIV YALE/
SAUNDERS M
WO 9606391 A3 US 5025388 A A 91-200832/27 (CRAM/) CRAMER R D/CRAMER
R D; WOLD S V
WO 9606391 A3 US 5056035 A A 91-317888/43 (FUJF ) FUJI PHOTO FILM
CO LTD/FUJITA S
CITED LITERATURE REFERENCES
Family Member Cat Citation
By Inventor:
WO 9606391 A2 Oracle Relational Database Management System by
Oracle Corporation, World Headquarters, 500 Oracle
Pkwy., Redwood Shores, CA 94065
WO 9606391 A2 Molecular Access System (MACCS) created by Molecular
Design Ltd., MDL Information Systems, 14600 Catalina
Street, San Leandro, CA 94577
WO 9606391 A2 Integrated Scientific Information System (ISIS)
created by Molecular Design Ltd., MDL Information
Systems, 14600 Catalina Street, San Leandro, CA
94577
EP 777882 A JOURNAL OF CHEMOMETRICS, ENGLAND GB, XP002040815
EP 777882 A See also references of WO 9606391A3
US 5577239 A Viking Instruments Corp. (Hewlett Packard); Spectra
Trak Transportable GC/MS System; (brochure), No
date.
US 5577239 A Chemical Structure, The International Language of
Chemistry; Wendy A. War (Ed.); "Interfacing
DARC-Oracle" AJCM (Juus) de Jong (1988).
US 5577239 A J. Chem. Inf. Comput. Sci. (1983), vol. 23, No. 3
pp. 102-108; DARC Substructure Search System; A New
Approach to Chemical Information; Roger Attias.
US 5577239 A J. Chem. Inf. Comput. Sci. (1987), vol. 27, No. 2;
pp. 74-82; DARC System; Notions of Defined and
Generic Substructures. Filiation and Coding of FREL
Substructure (SS) Classes; Jacques-Emile Dubois et
al.
US 5577239 A J. Chem. Inf. Comput. Sci. (1990), vol. 30, No. 2;
pp. 191-199, Substructure Search Systems, 1,
Performance Comparison of the MACCS, DARC, HTSS, CAS
Registry MVSSS, and S4 Substructure Search Systems;
Martin G. Hicks.
US 5577239 A J. Chem. Inf. Comput. Sci. (1988), vol. 28, No. 4;
pp. 221-226; An Efficient Graph Approach to Matching
Chemical Structures, O. Owolabi.
US 5577239 A J. Chem. Inf. Comput. Sci. (1990), vol. 30, No. 4;
pp. 332-339; Reactions in the Bellstein Information
System: Nonaporic Organic Synthesis; Martin G.
Hicks.
US 5577239 A Analytica Chimica Acta, 235 (1990), pp. 87-92;
Substructure Search Systems for Large Chemical Data
Bases; Martin G. Hicks et al.
US 5577239 A J. Chem. Inf. Comput. Sci. (1991), vol. 31, No. 2;
pp. 320-326; The Bellstein Structure Registry
System, 1, General Design; Laszio Domokos.
US 5577239 A J. Chem. Inf. Comput. Sci. (1989), vol. 29, No. 4;
pp. 255-260; 3DSearch; A System for
Three-Dimensional Substructure Searching; Robert P.
Sheridan, et al.
US 5577239 A Substructure Searches of Chemical Structure Files;
(Jan. 23, 1973); Strategic Considerations in the
Design of a Screening System for Substructure
Searches of Chemical Structure Files; George W.
Adamson, et al.
US 5577239 A Chemical Structure Searching; (Jan. 21, 1975); An
Efficient Design for Chemical Structure Searching,
I, The Screens; Alfred Feldman et al.
US 5577239 A J. Chem. Inf. Comput. Sci. (1982), vol. 22, No. 4;
The Third BASIC Fragment Search Dictionary; W. Graf,
H. K. Kaindl, et al.
US 5577239 A J. Chem. Inf. Comput. Sci. (1983), vol. 23, No. 3;
The CAS ONLINE Search System, 1, General System
Design and Selection, Generation, and Use of Search
Screens; P. G. Dittmar, et al.
US 5577239 A Computer Chemical, (1991), vol. 15, No. 2; pp.
103-107; A Central Atom Based Algorithm and Computer
Program for Substructure Search; Alf Dengler and
Ivar Ugi.
US 5577239 A J. Chem. Inf. Comput. Sci. (1993), vol. 33, No. 4;
pp. 545-547; Structure Searching in Chemical
Databases by Direct Lookup Methods; Bradley D.
Christie et al.
US 5577239 A J. Chem. Inf. Comput. Sci. (1993); vol. 33, No. 4;
pp. 539-541; Substructure Searching on Very Large
Files by Using Multiple Storage Techniques;
Alexander Bartmann et al.
US 5950192 A Viking Instruments Corp. (Hewlett Packard);
SpectraTrak Transportable GS/MS Systems;
(brochure)-No Date.
US 5950192 A Chemical Structures, The International Language of
Chemistry; Wendy A. War (Ed.); "Interfacing
DARC-Oracle" AJCM (Juus) de Jong (1988).
US 5950192 A J. Chem. Inf. Comput. Sci. (1983) , vol. 23, No. 3;
pp. 102-108; DARC Substructure Search System: A New
Approach to Chemical Information; Roger Attias.
US 5950192 A J. Chem. Inf. Comput. Sci. (1987), vol. 27, No. 2;
pp. 74-82; DARC System: Notions of Defined and
Generic Substructures. Filiation and Coding of FREL
Substructure (SS) Classes; Jacques-Emile Dubois et
al.
US 5950192 A J. Chem. Inf. Comput. Sci. (1990), vol. 30, No. 2;
pp. 191-199, Substructure Search Systems. 1.
Performance Comparison of the MACCS, DARC, HTSS, and
CAS Registry MVSSS, and S4 Substructure Search
System; Martin G. Hicks & Clemens.
US 5950192 A J. Chem. Inf. Comput. sci. (1988), vol. 28, No. 4;
pp. 221-226; An Efficient Graph Approach to Matching
Chemical Structures, O. Owolabi.
US 5950192 A J. Chem. Inf. Comput. Sci. (1990), vol. 30, No. 4;
pp. 332-339; Reactions in the Beilstein Information
System: Nonaporic Organic Synthesis; Martin G.
Hicks.
US 5950192 A Analytica Chimica Acta, 235 (1990), pp. 87-92;
Substructure Search Systems for Large Chemical Data
Bases; Martin G. Hicks et al.
US 5950192 A J. Chem. Inf. Comput. Sci. (1991), vol. 31, No. 2;
pp. 320-326; The Beilstein Structure Registry
System. 1. General Design; Laszio Domokos.
US 5950192 A J. Chem. Inf. Comput. Sci. (1989), vol. 29, No. 4;
pp. 255-260; 3DSearch; A System for
Three-Dimensional Substructure Searching; Robert P.
Sheridan, et al.
US 5950192 A Substructure Searches of Chemical Structure Files;
(Jan. 23, 1973); Strategic Considerations in the
Design of a Screening System for Substructure
Searches of Chemical Structure Files; George W.
Adamson, et al.
US 5950192 A Chemical Structure Searching; (Jan. 21, 1975); An
Efficient Design for Chemical Structure Searching.
I. The Screens; Alfred Feldman et al.
US 5950192 A J. Chem. Inf. Comput. Sci. (1982), vol. No. 4; The
Third BASIC Fragment Search Dictionary; W. Graf, H.
K. Kaindl, et al.
US 5950192 A J. Chem. Inf. Comput. Sci. (1983), vol. 23, No. 3;
The CAS Online Search System. 1. General System
Design and Selection, Generation, and Use of Search
Screens; P. G. Dittmar, et al.
US 5950192 A Computer Chemical, ((1991), vol. 15, No. 2, pp.
103-107; A Central Atom Based Algorithm and Computer
Program for Substructure Search; Alf Dengler and
Ivar Ugi.
US 5950192 A J. Chem. Inf. Comput. Sci. (1993), vol. 33, No. 4;
pp. 545-547; Sturcture Searching in Chemical
Databases by Direct Lookup Methods; Baradley D.
Christie et al.
WO 9606391 A See also references of EP 0777882A4
2/2/2
DIALOG(R)File 342:Derwent Patents Citation Indx
(c) 2001 Derwent Info Ltd. All rts. reserv.
00376357 WPI Acc No: 87-056518/08
Markush structure database system which can handle Markush queries - in
which separate specific atom and generic term connection tables are linked
to reference data e.g. patent nuMbers
Patent Assignee: (AMCH-) AMER CHEMICAL SOC
Author (Inventor): FISANICK W
Patent Family:
Patent No Kind Date Examiner Field of Search
US 4642762 A 870210 (BASIC)
Derwent Week (Basic): 8708
Priority Data: US 614219 (840525)
Applications: US 614219 (840525)
Derwent Class: J04; T01
Int Pat Class: G06F-015/40
Number of Patents: 001
Number of Countries: 001
Number of Cited Patents: 001
Number of Cited Literature References: 000
Number of Citing Patents: 026
CITED PATENTS
Family Member Cited Patent Cat WPI Acc No Assignee/Inventor
By Examiner:
US 4642762 A US 4473890 A 83-790076/42 (NIIN-) JAPAN INFORM CENT
/ARAKI K
?
S PN=US 5917146
S3 1 PN=US 5917146
?
TYPE 3/2/1
3/2/1
DIALOG(R)File 342:Derwent Patents Citation Indx
(c) 2001 Derwent Info Ltd. All rts. reserv.
03378757 WPI Acc No: 99-059790/05
Low-smoke pyrotechnic compositions - containing dihydrazino-s-tetrazine or
its derivatives or salts, an oxidizing agent and a colourant, useful for
firework displays and special effects in the film industry
Patent Assignee: (REGC ) UNIV CALIFORNIA; (HISK/) HISKEY M A; (CHAV/)
CHAVEZ D E
Author (Inventor): HISKEY M A; CHAVEZ D E
Patent Family:
Patent No Kind Date Examiner Field of Search
WO 9854113 A1 981203 (BASIC)
AU 9877092 A 981230
US 5917146 A 990629 149/36; 149/46; 149/61; 149/76
Derwent Week (Basic): 9905
Priority Data: US 865412 (970529)
Applications: US 865412 (970529); AU 9877092 (980529); WO 98US11062 (
980529)
Designated States
(National): AL; AM; AT; AU; AZ; BA; BB; BG; BR; BY; CA; CH; CN; CU; CZ;
DE; DK; EE; ES; FI; GB; GE; GH; GM; GW; HU; ID; IL; IS; JP; KE; KG; KP
; KR; KZ; LC; LK; LR; LS; LT; LU; LV; MD; MG; MK; MN; MW; MX; NO; NZ;
PL; PT; RO; RU; SD; SE; SG; SI; SK; SL; TJ; TM; TR; TT; UA; UG; UZ; VN
; YU; ZW
(Regional): AT; BE; CH; CY; DE; DK; EA; ES; FI; FR; GB; GH; GM; GR; IE;
IT; KE; LS; LU; MC; MW; NL; OA; PT; SD; SE; SZ; UG; ZW
Derwent Class: E13; K04
Int Pat Class: C06B-031/02
Number of Patents: 003
Number of Countries: 082
Number of Cited Patents: 016
Number of Cited Literature References: 001
Number of Citing Patents: 001
CITED PATENTS
Family Member Cited Patent Cat WPI Acc No Assignee/Inventor
By Examiner:
US 5917146 A US 3244702 A /MARCUS
US 5917146 A US 3697339 A 71-68351S/43 (MESR )
MESSERSCHMITT-BOELKOW-BLOHM GMBH
US 5917146 A US 3797238 A 74-24569V/13 (UNAC ) UNITED AIRCRAFT
CORP
US 5917146 A US 3940298 A 76-18661X/10 (USNA ) US SEC OF NAVY
US 5917146 A US 4078954 A 77-06100Y/04 (POUE ) SOC NAT POUDRES &
EXPLOSIFS
US 5917146 A US 5197758 A 93-125801/15 (MORN ) MORTON INT INC/
LUND G K; STEVENS M R; EDWARDS W W; SHAW G C
US 5917146 A US 5198046 A 92-115569/15 (FRAU ) FRAUNHOFER-GES
FORD ANGE/BUCERIUS K M; WASMANN F W; MENKE K
US 5917146 A US 5281706 A 94-042864/05 (USAT ) US DEPT ENERGY/
OTT D G; COBURN M D
US 5917146 A US 5449423 A 94-151167/18 (CIOF/) CIOFFE A/CIOFFE A
US 5917146 A US 5472534 A 96-029697/03 (THIO ) THIOKOL CORP/
WARDLE R B; BLAU R J
US 5917146 A US 5525166 A 94-237199/29 (STFI-) STANDARD
FIREWORKS LTD/COOK B
WO 9854113 A US 3797238 A A 74-24569V/13 (UNAC ) UNITED AIRCRAFT
CORP
WO 9854113 A US 3940298 A A 76-18661X/10 (USNA ) US SEC OF NAVY
WO 9854113 A US 4078954 A A 77-06100Y/04 (POUE ) SOC NAT POUDRES &
EXPLOSIFS
WO 9854113 A US 5449423 A A 94-151167/18 (CIOF/) CIOFFE A/CIOFFE A
WO 9854113 A US 5525166 A A 94-237199/29 (STFI-) STANDARD
FIREWORKS LTD/COOK B
CITED LITERATURE REFERENCES
Family Member Cat Citation
By Examiner:
US 5917146 A Marcus et al., Journal of Org. Chem., vol. 28, pp.
2372-2378, Sep. 1963.
?
S FLUOXETINE
S4 51 FLUOXETINE
?
TYPE 4/TI/1-5
4/TI/1
DIALOG(R)File 342:(c) 2001 Derwent Info Ltd. All rts. reserv.
Crystallizing (R)-fluoxetine hydrochloride from a mixed solvent in high
enantiomeric purity with high recovery rate ...
4/TI/2
DIALOG(R)File 342:(c) 2001 Derwent Info Ltd. All rts. reserv.
Preparation of (S)-3-phenyl-3-phenoxy propylamine derivatives e.g.
(S)-fluoxetine, comprises converting 3-substituted-1-phenyl propene
derivative to racemic epoxide or alcohol, resolving and further reacting
pure enantiomer...
4/TI/3
DIALOG(R)File 342:(c) 2001 Derwent Info Ltd. All rts. reserv.
Production of fluoxetine, useful as antidepressant, from
3-halo-1-phenylpropan-1-ol without use of protecting groups...
4/TI/4
DIALOG(R)File 342:(c) 2001 Derwent Info Ltd. All rts. reserv.
Composition comprising R(-)-fluoxetine, mostly free of the S(+)-isomer, and
9-hydroxy-risperidone, for the treatment of psychotic or psychiatric
disorders in children, e.g. anxiety disorder, depression, Tourette's
disorder...
4/TI/5
DIALOG(R)File 342:(c) 2001 Derwent Info Ltd. All rts. reserv.
N,N-dimethyl-(3-phenyl-3-((4-trifluoromethyl)-phenoxy)-propylamine)-p-tolue
ne sulfonate is a new intermediate for the synthesis of fluoxetine...
?
S S4 AND LILLY
51 S4
0 LILLY
S5 0 S4 AND LILLY
?
S S4 AND PA=LILLY
51 S4
3949 PA=LILLY
S6 18 S4 AND PA=LILLY
?
TYPE 6/2/1-5
6/2/1
DIALOG(R)File 342:Derwent Patents Citation Indx
(c) 2001 Derwent Info Ltd. All rts. reserv.
04371800 WPI Acc No: 01-031829/04
Crystallizing (R)-fluoxetine hydrochloride from a mixed solvent in high
enantiomeric purity with high recovery rate -
Patent Assignee: (ELIL ) LILLY & CO ELI
Author (Inventor): BRENNAN J; DISEROAD W D; HAY L A; MITCHELL D
Patent Family:
Patent No Kind Date Examiner Field of Search
WO 200068182 A1 001116 (BASIC)
AU 200047984 A 001121
Derwent Week (Basic): 0104
Priority Data: US 133264 (990510)
Applications: AU 200047984 (000428); WO 2000US9801 (000428)
Designated States
(National): AE; AG; AL; AM; AT; AU; AZ; BA; BB; BG; BR; BY; CA; CH; CN;
CR; CU; CZ; DE; DK; DM; DZ; EE; ES; FI; GB; GD; GE; GH; GM; HR; HU; ID
; IL; IN; IS; JP; KE; KG; KP; KR; KZ; LC; LK; LR; LS; LT; LU; LV; MA;
MD; MG; MK; MN; MW; MX; NO; NZ; PL; PT; RO; RU; SD; SE; SG; SI; SK; SL
; TJ; TM; TR; TT; TZ; UA; UG; US; UZ; VN; YU; ZA; ZW
(Regional): AT; BE; CH; CY; DE; DK; EA; ES; FI; FR; GB; GH; GM; GR; IE;
IT; KE; LS; LU; MC; MW; NL; OA; PT; SD; SE; SL; SZ; TZ; UG; ZW
Derwent Class: B05
Int Pat Class: C07C-213/10
Number of Patents: 002
Number of Countries: 092
Number of Cited Patents: 001
Number of Cited Literature References: 000
Number of Citing Patents: 000
CITED PATENTS
Family Member Cited Patent Cat WPI Acc No Assignee/Inventor
By Examiner:
WO 200068182A EP 545425 A A 93-184173/23 (HMRI ) HOECHST ROUSSEL
PHARM INC/EFFLAND R C; KLEIN J T
6/2/2
DIALOG(R)File 342:Derwent Patents Citation Indx
(c) 2001 Derwent Info Ltd. All rts. reserv.
04208425 WPI Acc No: 01-015842/02
Process for preparing racemic fluoxetine, used for treatment of depression,
from mixture enriched in either enantiomer comprises reacting the mixture
with base in aprotic highly dipolar solvent -
Patent Assignee: (ELIL ) LILLY & CO ELI
Author (Inventor): KOENIG T M; MITCHELL D
Patent Family:
Patent No Kind Date Examiner Field of Search
WO 200064855 A1 001102 (BASIC)
AU 200040100 A 001110
Derwent Week (Basic): 0102
Priority Data: US 131074 (990426)
Applications: AU 200040100 (000328); WO 2000US6683 (000328)
Designated States
(National): AE; AG; AL; AM; AT; AU; AZ; BA; BB; BG; BR; BY; CA; CH; CN;
CR; CU; CZ; DE; DK; DM; DZ; EE; ES; FI; GB; GD; GE; GH; GM; HR; HU; ID
; IL; IN; IS; JP; KE; KG; KP; KR; KZ; LC; LK; LR; LS; LT; LU; LV; MA;
MD; MG; MK; MN; MW; MX; NO; NZ; PL; PT; RO; RU; SD; SE; SG; SI; SK; SL
; TJ; TM; TR; TT; TZ; UA; UG; US; UZ; VN; YU; ZA; ZW
(Regional): AT; BE; CH; CY; DE; DK; EA; ES; FI; FR; GB; GH; GM; GR; IE;
IT; KE; LS; LU; MC; MW; NL; OA; PT; SD; SE; SL; SZ; TZ; UG; ZW
Derwent Class: B05
Int Pat Class: C07B-055/00
Number of Patents: 002
Number of Countries: 092
Number of Cited Patents: 001
Number of Cited Literature References: 000
Number of Citing Patents: 000
CITED PATENTS
Family Member Cited Patent Cat WPI Acc No Assignee/Inventor
By Examiner:
WO 200064855A US 5847214 A A 99-059172/05 (LAPO-) LAPORTE ORGANICS
SPA FRANCIS/ROSSETTI V; BERATTO S G V; AROSIO R
6/2/3
DIALOG(R)File 342:Derwent Patents Citation Indx
(c) 2001 Derwent Info Ltd. All rts. reserv.
03268666 WPI Acc No: 98-586218/50
Fluoxetine enteric pellet - comprising fluoxetine and excipients, optional
separating layer, enteric layer of hydroxypropyl methylcellulose acetate
succinate and excipients and optional finishing layer
Patent Assignee: (ELIL ) LILLY & CO ELI
Author (Inventor): ANDERSON N R; HARRISON R G; LYNCH D F; OREN P L
Patent Family:
Patent No Kind Date Examiner Field of Search
GB 2325623 A 981202 (BASIC) None
AT 408068 B 010715
AT 9800931 A 010115
AU 726690 B 001116
AU 9869048 A 981203
BE 1011925 A3 000307
BR 9801989 A 000208
CA 2234826 A 981129
CA 2234826 C 001219
CN 1200924 A 981209
CN 1285189 A 010228
CZ 9801143 A3 981216
DE 19823940 A1 981203 A61K-009/6; A61K-031/135
DK 9800728 A 981130
FI 9800846 A 981130
FR 2763846 A1 981204
GB 2325623 B 990414 None
HU 9800882 A2 000328
JP 10330253 A 981215
KR 98086622 A 981205
MX 9803636 A1 990201
NL 1009259 C2 981201
NO 9802197 A 981130
NZ 330192 A 990828
PT 102152 A 981231
RU 2164405 C2 010327
SE 9801336 A 981130
SG 72805 A1 000523
US 5910319 A 990608 424/458; 424/459; 424/461; 424/464; 424/465;
424/489; 514/646; 514/962
US 5985322 A 991116 424/458; 424/459; 424/461; 424/464; 424/465;
424/489; 424/490; 424/494; 514/646; 514/962
ZA 9803173 A 991229
Derwent Week (Basic): 9850
Priority Data: US 867196 (970529)
Applications: US 867196 (970529); CA 2234826 (980414); RU 98106998 (980414
); CZ 981143 (980415); GB 987939 (980415); HU 98882 (980415); NZ 330192
(980415); ZA 983173 (980415); FI 98846 (980416); SE 981336 (980417); SG
98871 (980417); KR 9814014 (980420); CN 98108778 (980424); JP 98153501
(980424); CN 2000122209 (980424); MX 3636 (980507); PT 102152 (980508);
FR 986040 (980513); NO 982197 (980514); BR 981989 (980519); BE 98383 (
980520); NL 981009259 (980526); AU 9869048 (980528); DE 19823940 (
980528); DK 98728 (980528); AT 98931 (980529); US 265610 (990310)
Derwent Class: A96; B05
Int Pat Class: A61K-009/16; A61K-009/20; A61K-009/24; A61K-009/28;
A61K-009/30; A61K-009/32; A61K-009/36; A61K-009/48; A61K-009/50;
A61K-009/52; A61K-009/54; A61K-009/60; A61K-009/62; A61K-031/02;
A61K-031/13; A61K-031/135; A61K-031/138; A61K-031/40; A61K-047/36;
A61K-047/38; A61P-025/24
Number of Patents: 031
Number of Countries: 025
Number of Cited Patents: 034
Number of Cited Literature References: 039
Number of Citing Patents: 000
CITED PATENTS
Family Member Cited Patent Cat WPI Acc No Assignee/Inventor
By Examiner:
GB 2325623 A EP 687472 A2 96-031591/04 (ELIL ) LILLY & CO ELI/
WONG D T; OGUIZA J I
GB 2325623 A US 4847092 A 87-137609/20 (ELIL ) LILLY & CO ELI/
THAKKAR A L; GIBSON L L
GB 2325623 A US 5508276 A 96-078389/09 (ELIL ) LILLY & CO ELI;
(SHIO ) SHIONOGI & CO LTD/ANDERSON N R; OREN P L;
OGURA T; FUJII T
GB 2325623 B EP 687472 A2 96-031591/04 (ELIL ) LILLY & CO ELI/
WONG D T; OGUIZA J I
GB 2325623 B US 4847092 A 87-137609/20 (ELIL ) LILLY & CO ELI/
THAKKAR A L; GIBSON L L
GB 2325623 B US 5508276 A 96-078389/09 (ELIL ) LILLY & CO ELI;
(SHIO ) SHIONOGI & CO LTD/ANDERSON N R; OREN P L;
OGURA T; FUJII T
US 5910319 A EP 687472 A2 96-031591/04 (ELIL ) LILLY & CO ELI/
WONG D T; OGUIZA J I
US 5910319 A EP 693281 A2 96-078388/09 (ELIL ) LILLY SA/ARCE
MENDIZABAL F
US 5910319 A US 4314081 A 75-49665W/30 (ELIL ) LILLY & CO ELI
US 5910319 A US 4444778 A 84-120590/19 (COUG/) COUGHLIN S R/
COUGHLIN S R
US 5910319 A US 4626549 A 86-338922/51 (ELIL ) LILLY & CO ELI/
MOLLOY B B; SCHMIEGEL K K
US 5910319 A US 4847092 A 87-137609/20 (ELIL ) LILLY & CO ELI/
THAKKAR A L; GIBSON L L
US 5910319 A US 5104899 A 92-150223/18 (SEPR-) SEPRACOR INC/
YOUNG J W; BARBERICH T J
US 5910319 A US 5356934 A 91-289974/40 (ELIL ) LILLY & CO ELI/
ROBERTSON D W; WONG D T
US 5910319 A US 5508276 A 96-078389/09 (ELIL ) LILLY & CO ELI;
(SHIO ) SHIONOGI & CO LTD/ANDERSON N R; OREN P L;
OGURA T; FUJII T
US 5910319 A WO 9213452 A1 92-299671/36 (YOUN/) YOUNG J W;
(BARB/) BARBERICH T J; (TEIC/) TEICHER M H/YOUNG
J W; BARBERICH T J; TEICHER M H
US 5910319 A WO 9219226 A1 92-398501/48 (DYNA-) DYNAGEN INC/
KITCHELL J P; MUNI I A; BOYER Y N
US 5910319 A WO 9318755 A1 93-320425/40 (DEPO-) DEPOMED SYSTEMS
INC/SHELL J W
US 5910319 A WO 9324154 A1 93-405434/50 (FUIS-) FUISZ
TECHNOLOGIES LTD/FUISZ R C
US 5910319 A WO 9512385 A1 95-185578/24 (ISOT-) ISOTECH MEDICAL
INC/CHO Y W
US 5985322 A EP 687472 A2 96-031591/04 (ELIL ) LILLY & CO ELI/
WONG D T; OGUIZA J I
US 5985322 A EP 693281 A2 96-078388/09 (ELIL ) LILLY SA/ARCE
MENDIZABAL F
US 5985322 A US 4314081 A 75-49665W/30 (ELIL ) LILLY & CO ELI
US 5985322 A US 4444778 A 84-120590/19 (COUG/) COUGHLIN S R/
COUGHLIN S R
US 5985322 A US 4626549 A 86-338922/51 (ELIL ) LILLY & CO ELI/
MOLLOY B B; SCHMIEGEL K K
US 5985322 A US 4847092 A 87-137609/20 (ELIL ) LILLY & CO ELI/
THAKKAR A L; GIBSON L L
US 5985322 A US 5104899 A 92-150223/18 (SEPR-) SEPRACOR INC/
YOUNG J W; BARBERICH T J
US 5985322 A US 5356934 A 91-289974/40 (ELIL ) LILLY & CO ELI/
ROBERTSON D W; WONG D T
US 5985322 A US 5508276 A 96-078389/09 (ELIL ) LILLY & CO ELI;
(SHIO ) SHIONOGI & CO LTD/ANDERSON N R; OREN P L;
OGURA T; FUJII T
US 5985322 A WO 9213452 A1 92-299671/36 (YOUN/) YOUNG J W;
(BARB/) BARBERICH T J; (TEIC/) TEICHER M H/YOUNG
J W; BARBERICH T J; TEICHER M H
US 5985322 A WO 9219226 A1 92-398501/48 (DYNA-) DYNAGEN INC/
KITCHELL J P; MUNI I A; BOYER Y N
US 5985322 A WO 9318755 A1 93-320425/40 (DEPO-) DEPOMED SYSTEMS
INC/SHELL J W
US 5985322 A WO 9324154 A1 93-405434/50 (FUIS-) FUISZ
TECHNOLOGIES LTD/FUISZ R C
US 5985322 A WO 9512385 A1 95-185578/24 (ISOT-) ISOTECH MEDICAL
INC/CHO Y W
CITED LITERATURE REFERENCES
Family Member Cat Citation
By Examiner:
US 5910319 A Johnson, World Patents Index, #93-352132, 1993.
US 5910319 A Oguiza et al., World Patents Index, #96-031591,
1996.
US 5910319 A Wong et al., Chemical Abstracts, vol. 124, #156011,
1995.
US 5910319 A Montgomery, et al., Eur. Arch. Psychiatry Clin.
Neurosci., 244:211-215 (1994).
US 5910319 A Burke, et al., Psychopharmacol. Bull. 31(3); 524
(1995).
US 5910319 A Stafford, et al., Drug Development and Industrial
Pharmacy, 8(4):513-530 (1982).
US 5910319 A Osterwald, Hermann P., Pharmaceutical Research,
2:14-18 (1985).
US 5910319 A Davis, et al., Drug Development and Industrial
Pharmacy, 12(10):1419-1448 (1986).
US 5910319 A Bloor, et al., Drug Development and Industrial
Pharmacy, 15(15-16):2227-2243.
US 5910319 A Nagai, et al., Aqueous Polymeric Coating for
Pharmaceutical Dosage Forms, Marcel Dekker, N.W. and
Basel, 81-152 (1989).
US 5910319 A Chang, Rong-Kun, Pharmaceutical Technology,
14(10):2-70 (1990).
US 5910319 A Fujii, et al., Recent Advances On Aqueous Polymeric
Coating System and Related Techniques, Proceedings
of Pre-World Congress Particle Technology in Gifu,
Sep. 17-18, 1990, Gifu, Japan, 80-85 (1990).
US 5910319 A Delattre, et al., Proceed. Intern. Symp. Control.
Rel. Bioact. Mater., 19:267-268 (1992).
US 5910319 A Schmidt, et al., Drug Development and Industrial
Pharmacy, 18(18):1969-1979 (1992).
US 5910319 A Wyatt "Enhanced Stability of Aqueous Cellulose
Acetate Phthalate (CAP) Enteric Films." Presented
AAPS Annual Meeting, San Antonio, TX, Nov. 15-19,
1992.
US 5910319 A Takahata, et al., Chemical and Pharmaceutical
Bulletin, 41(6):1137-1143 (1993).
US 5910319 A Obara, et al., Pharmaceutical Research,
11(11):1562-1567 (1994).
US 5910319 A Shin-Etsu Chemical Co., Ltd., "An Improved Aqueous
Coating Using Shin-Etsu AQOAT", AQOAT Technical
Information Bulletin, 1994.
US 5910319 A Japan Pharmaceutical Excipients Council,
"Hydroxypropylmethylcellulose Acetate Succinate",
Japanese Pharmaceutical Excipients 1993 (JPE 1993),
183-187, Yakuji Nippo, Ltd., Tokyo, Japan, 1994.
US 5910319 A Shin-Etsu Chemical Co., Ltd. "Dry Coating", AQOAT
Technical Inormation No. A-3, Sep., 1996.
US 5910319 A Obara, et al., Poster PT6115, Dry Coating'-A Novel.
US 5985322 A Montgomery, et al., Eur. Arch. Psychiatry Clin.
Neurosci., 244:211-215 (1994).
US 5985322 A Burke, et al., Psychopharmacol. Bull. 31 (3); 524
(1995).
US 5985322 A Stafford, et al., Drug Development and Industrial
Pharmacy, 8(4) :513-530 (1982).
US 5985322 A Osterwald, Hermann P., Pharmaceutical Research,
2:14-18 (1985).
US 5985322 A Davis, et al., Drug Development and Industrial
Pharmacy, 12(10):1419-1448 (1986).
US 5985322 A Bloor, et al., Drug Development and Industrial
Pharmacy, 15 (15-16) :2227-2243.
US 5985322 A Nagai, et al., Aqueous Polymeric Coating for
Pharmaceutical Dosage Forms, Marcel Dekker, N.W. and
Basel, 81-152 (1989).
US 5985322 A Chang, Rong-Kun, Pharmaceutical Technology, 14(10)
:2-70 (1990).
US 5985322 A Fujii, et al., Recent Advances On Aqueous Polymeric
Coating System and Related Techniques, Proceedings
of Pre-World Congress Particle Technology in Gifu,
Sep. 17-18, 1990, Gifu,Japan, 80-85 (1990).
US 5985322 A Delattre, et al., Proceed. Intern. Symp. Control.
Rel. Bioact. Mater., 19:267-268 (1992).
US 5985322 A Schmidt, et al., Drug Development and Industrial
Pharmacy, 18(18) :1969-1979 (1992).
US 5985322 A Wyatt "Enhanced Stability of Aqueous Cellulose
Acetate Phthalate (CAP) Enteric Films." Presented
AAPS Annual Meeting, San Antonio, TX, Nov. 15-19,
1992.
US 5985322 A Takahata, et al., Chemical and Pharmaceutical
Bulletin, 41(6) :1137-1143 (1993).
US 5985322 A Obara, et al., Pharmaceutical Research, 11(11)
:1562-1567 (1994).
US 5985322 A Shin-Etsu Chemical CO., Ltd., "An Improved Aqueous
Coating using Shin-Etsu AQOAT", AQOAT Technical
Information Bulletin, 1994.
US 5985322 A Japan Pharmaceutical Excipients Council,
"Hydroxypropylmethylcellulose Acetate Succinate",
Japanese Pharmaceutical Excipients 1993 (JPE 1993),
183-187, Yakuji Nippo, Ltd., Tokyo, Japan, 1994.
US 5985322 A Shin-Etsu Chemical Co., Ltd. "Dry Coating", AQOAT
Technical Inormation No. A-3, Sep., 1996.
US 5985322 A Obara, et al., Poster PT6115, Dry Coating-A Novel.
6/2/4
DIALOG(R)File 342:Derwent Patents Citation Indx
(c) 2001 Derwent Info Ltd. All rts. reserv.
03015930 WPI Acc No: 98-065080/07
Potentiation of anti-depressant drugs e.g. fluoxetine, using
N-(2-(4-(2-methoxyphenyl)piperazin-1-yl))-N-(2-pyridyl)cyclohexanecarboxami
de - increases serotonin levels in patients beyond levels normally achieved
with drug alone
Patent Assignee: (ELIL ) LILLY SA
Author (Inventor): ARTIGAS PEREZ F
Patent Family:
Patent No Kind Date Examiner Field of Search
EP 818198 A1 980114 (BASIC)
Derwent Week (Basic): 9807
Priority Data: EP 96500097 (960709)
Applications: EP 96500097 (960709)
Designated States
(Regional): ES
Derwent Class: B03; B05
Int Pat Class: A61K-031/505; A61K-031-135
Number of Patents: 001
Number of Countries: 001
Number of Cited Patents: 003
Number of Cited Literature References: 009
Number of Citing Patents: 000
CITED PATENTS
Family Member Cited Patent Cat WPI Acc No Assignee/Inventor
By Examiner:
EP 818198 A EP 687472 A Y 96-031591/04 (ELIL ) LILLY & CO ELI/
WONG D T; OGUIZA J I
EP 818198 A EP 714663 A Y 96-261402/27 (ELIL ) LILLY & CO ELI/
OGUIZA J I; WONG D T
EP 818198 A US 4940585 A A 90-231154/30 (HAPW/) HAPWORTH W E/
HAPWORTH W E; HAPWORTH M S
CITED LITERATURE REFERENCES
Family Member Cat Citation
By Examiner:
EP 818198 A BR. J. PHARMACOL., vol. 115, no. 6, 1995, pages
1064-1070, XP000604130 GARTSIDE ET AL: "INTERACTION
BETWEEN A SELECTIVE 5-HT1A RECEPTOR ANTAGONIST AND
AN SSRI IN VIVO: EFFECTS ON 5-HT CELL FIRING AND
EXTRACELLULAR 5-HT"
EP 818198 A ARCHIVES OF GENERAL PSYCHIATRY, vol. 51, no. 3,
1994, pages 248-251, XP000605678 ARTIGAS ET AL:
"PINDOLOL INDUCES A RAPID IMPROVEMENT OF DEPRESSED
PATIENTS TREATED WITH SEROTONIN REUPTAKE INHIBITORS"
EP 818198 A J. NEUROCHEM., vol. 66, no. 2, 1996, pages 599-603,
XP000604121 ENGLEMAN ET AL: "ANTAGONISM OF SEROTONIN
5-HT1A RECEPTORS POTENTIATES THE INCREASES IN
EXTRACELLULAR MONOAMINES INDUCED BY DULOXETINE IN
RAT HYPOTHALAMUS"
EP 818198 A Newman-Tancredi, A. et al.: NEUROPHARMACOLOGY, vol.
36, no. 4/5, 1997, pp.451-459
EP 818198 A Newman-Tancredi, A. et al.: NEUROPHYCHOPHARMACOLOGY,
vol. 18, no,. 5 1998pp.395-398
EP 818198 A Romero, L. et al.: Strategies to optimize the
antidepressant action of selective serotonin
reuptake inhibitors in ANTIDEPRESSANTS: NEW
PHARMACOLOGICAL STRATEGIES,1996, Ed. P. Skolnic, pp.
1-33, Humana Press, Totowa, NJ, USA
EP 818198 A Romero, L. et al.: NEUROPSYCHOPHARMACOLOGY, vol. 15,
no. 4, 1996, pp. 349-360
EP 818198 A Romero, L. et al.: JOURNAL OF NEUROCHEMISTRY, vol.
68, 1997, pp. 2593-2603
EP 818198 A Romero, L. et al.: NEUROSCIENCE LETTERS, vol. 219,
1996, pp. 123-126
6/2/5
DIALOG(R)File 342:Derwent Patents Citation Indx
(c) 2001 Derwent Info Ltd. All rts. reserv.
02847079 WPI Acc No: 97-427083/40
Treatment of sleep disorders with two of specified compounds - one a
serotonin uptake inhibitor and the other a serotonin receptor 1A
antagonist, e.g. fluoxetine or duloxetine, and pindolol
Patent Assignee: (ELIL ) LILLY & CO ELI
Author (Inventor): JAMES S P
Patent Family:
Patent No Kind Date Examiner Field of Search
EP 792649 A1 970903 (BASIC)
AU 9720586 A 970916
WO 9731629 A1 970904
Derwent Week (Basic): 9740
Priority Data: US 12523 (960229)
Applications: EP 97301057 (970219); AU 9720586 (970227); WO 97US3068 (
970227)
Designated States
(National): AL; AM; AU; AZ; BA; BB; BG; BR; BY; CA; CN; CU; CZ; EE; GE;
GH; HU; IL; IS; JP; KE; KG; KP; KR; KZ; LC; LK; LR; LS; LV; MD; MG; MK
; MN; MW; MX; NO; NZ; PL; RO; RU; SD; SG; SI; SK; TJ; TM; TR; TT; UA;
UG; US; UZ; YU
(Regional): AT; BE; CH; DE; DK; EA; ES; FI; FR; GB; GH; GR; IE; IT; KE;
LI; LS; LU; MW; NL; OA; PT; RO; SD; SE; SZ; UG
Derwent Class: B02; B05
Int Pat Class: A61K-031/05
Number of Patents: 003
Number of Countries: 074
Number of Cited Patents: 004
Number of Cited Literature References: 002
Number of Citing Patents: 001
CITED PATENTS
Family Member Cited Patent Cat WPI Acc No Assignee/Inventor
By Examiner:
EP 792649 A EP 687472 A X 96-031591/04 (ELIL ) LILLY & CO ELI/
WONG D T; OGUIZA J I
EP 792649 A US 4584404 A Y 86-125093/19 (ELIL ) LILLY & CO ELI/
MOLLOY B B; SCHMIEGEL K K
EP 792649 A US 5250571 A Y 91-261604/36 (ELIL ) LILLY & CO ELI/
FULLER R W; MITCHELL D; ROBERTSON D W; STEPHENSON
G A; WONG D T
EP 792649 A US 5356934 A Y 91-289974/40 (ELIL ) LILLY & CO ELI/
ROBERTSON D W; WONG D T
CITED LITERATURE REFERENCES
Family Member Cat Citation
By Examiner:
EP 792649 A ARCHIVES OF GENERAL PSYCHIATRY, vol. 51, no. 3,
March 1994, pages 248-251, XP000605678 "PINDOLOL
INDUCES A RAPID IMPROVEMENT OF DEPRESSED PATIENTS
TREATED WITH SEROTONIN REUPTAKE INHIBITORS"
WO 9731629 A DATABASE CA ON STN, No. 107:1364, HILAKIVI I. et
al., "Effects of Serotonin and Noradrenaline Uptake
Blockers on Wakefulness and Sleep in Cats"; &
PARMACOL. TOXICOLOGY, 60(3), 1987.
?
TYPE /2/15-18
6/2/15
DIALOG(R)File 342:Derwent Patents Citation Indx
(c) 2001 Derwent Info Ltd. All rts. reserv.
00541572 WPI Acc No: 88-347698/49
Antidiabetic fluoxetine compsn. - without causing major body weight loss
Patent Assignee: (ELIL ) LILLY & CO ELI
Author (Inventor): WONG D T
Patent Family:
Patent No Kind Date Examiner Field of Search
EP 294028 A 881207 (BASIC)
AU 8815533 A 881110
DE 3883606 G 931007
DK 8802385 A 881105
EP 294028 B1 930901
ES 2058272 T3 941101
JP 2721507 B2 980304 0 None
JP 63284126 A 881121
ZA 8803115 A 900131
Derwent Week (Basic): 8849
Priority Data: US 45509 (870504)
Applications: DE 3883606 (880429); EP 88303930 (880429); JP 88109838 (
880502)
Designated States
(Regional): AT; BE; CH; DE; ES; FR; GB; GR; IT; LI; LU; NL; SE
Derwent Class: B05
Int Pat Class: A61K-031/13; A61K-031/135
Number of Patents: 009
Number of Countries: 017
Number of Cited Patents: 003
Number of Cited Literature References: 000
Number of Citing Patents: 004
CITED PATENTS
Family Member Cited Patent Cat WPI Acc No Assignee/Inventor
By Examiner:
EP 294028 A US 4626549 A 86-338922/51 (ELIL ) LILLY & CO ELI/
MOLLOY B B; SCHMIEGEL K K
EP 294028 B1 US 4626549 A 86-338922/51 (ELIL ) LILLY & CO ELI/
MOLLOY B B; SCHMIEGEL K K
JP 2721507 B2 US 4626549 A 86-338922/51 (ELIL ) LILLY & CO ELI/
MOLLOY B B; SCHMIEGEL K K
6/2/16
DIALOG(R)File 342:Derwent Patents Citation Indx
(c) 2001 Derwent Info Ltd. All rts. reserv.
00334893 WPI Acc No: 86-233903/36
Analgesic compsn. contg. codeine and fluoxetine or norfluoxetine - and opt.
aspirin or acetominophen
Patent Assignee: (ELIL ) LILLY & CO ELI
Author (Inventor): HYNES M D
Patent Family:
Patent No Kind Date Examiner Field of Search
EP 193355 A 860903 (BASIC)
AU 8654001 A 860828
CA 1267092 A 900327
DE 3684626 G 920507
EP 193355 B 920401
JP 61200911 A 860905
JP 95045405 B2 950517
US 4683235 A 870728
ZA 8601211 A 870818
Derwent Week (Basic): 8636
Priority Data: US 705176 (850225); US 889157 (860725)
Applications: ZA 861211 (860218); EP 86301207 (860220); JP 8640123 (860224
); US 889157 (860725)
Designated States
(Regional): BE; CH; DE; FR; GB; IT; LI; LU; NL; SE
Derwent Class: B05
Int Pat Class: A61K-031/135; A61K-031/485; A61K-031/61; A61K-031/615
Number of Patents: 009
Number of Countries: 015
Number of Cited Patents: 010
Number of Cited Literature References: 005
Number of Citing Patents: 005
CITED PATENTS
Family Member Cited Patent Cat WPI Acc No Assignee/Inventor
By Examiner:
EP 193355 A FR 1373 A 66-05230F/00 (SOIF ) SOC IND
FABRICATION ANTIBIOTIQUES
EP 193355 A US 4035511 A 77-52098Y/29 (MASI ) MASSACHUSETTS
INST TECHNOLOGY
EP 193355 A US 4083982 A 78-40255A/22 (MASI ) MASSACHUSETTS
INST TECHNOLOGY/MESSING R B; LYTLE L D
EP 193355 B FR 1373 A 66-05230F/00 (SOIF ) SOC IND
FABRICATION ANTIBIOTIQUES
EP 193355 B US 4035511 A 77-52098Y/29 (MASI ) MASSACHUSETTS
INST TECHNOLOGY
EP 193355 B US 4083982 A 78-40255A/22 (MASI ) MASSACHUSETTS
INST TECHNOLOGY/MESSING R B; LYTLE L D
US 4683235 A US 4035511 A 77-52098Y/29 (MASI ) MASSACHUSETTS
INST TECHNOLOGY
US 4683235 A US 4083982 A 78-40255A/22 (MASI ) MASSACHUSETTS
INST TECHNOLOGY/MESSING R B; LYTLE L D
US 4683235 A US 4313896 A 82-13654E/07 (ELIL ) LILLY & CO ELI/
MOLLOY B B; SCHMIEGEL K K
US 4683235 A US 4314081 A 75-49665W/30 (ELIL ) LILLY & CO ELI
CITED LITERATURE REFERENCES
Family Member Cat Citation
By Examiner:
US 4683235 A Chem. Abst. 92 (1980) 191195z.
US 4683235 A Hynes et al., Drug Development Research, 2, 33
(1982).
US 4683235 A Messing et al., Psychopharmacology Communications,
1(5), 511 (1975).
US 4683235 A Larson et al., Life Sciences, 21, 1807 (1977).
US 4683235 A Sugrue et al., J. Pharm. Pharmac., 28, 447 (1976).
6/2/17
DIALOG(R)File 342:Derwent Patents Citation Indx
(c) 2001 Derwent Info Ltd. All rts. reserv.
00319969 WPI Acc No: 86-169186/26
Potentiation of dextropropoxyphene analgesia - using fluoxetine or
nor-fluoxetine
Patent Assignee: (ELIL ) LILLY & CO ELI
Author (Inventor): HYNES M D
Patent Family:
Patent No Kind Date Examiner Field of Search
US 4594358 A 860610 (BASIC)
AU 8653900 A 860828
CA 1268120 A 900424
EP 193354 A 860903
JP 61200912 A 860905
ZA 8601212 A 870818
Derwent Week (Basic): 8626
Priority Data: US 705177 (850225)
Applications: US 705177 (850225); ZA 861212 (860218); EP 86301206 (860220
); JP 8640124 (860224)
Designated States
(Regional): BE; CH; DE; FR; GB; IT; LI; LU; NL; SE
Derwent Class: B05
Int Pat Class: A61K-031/13
Number of Patents: 006
Number of Countries: 015
Number of Cited Patents: 007
Number of Cited Literature References: 005
Number of Citing Patents: 011
CITED PATENTS
Family Member Cited Patent Cat WPI Acc No Assignee/Inventor
By Examiner:
EP 193354 A US 4012525 A 77-21776Y/12 (ELIL ) LILLY & CO ELI
EP 193354 A US 4035511 A 77-52098Y/29 (MASI ) MASSACHUSETTS
INST TECHNOLOGY
EP 193354 A US 4083982 A 78-40255A/22 (MASI ) MASSACHUSETTS
INST TECHNOLOGY/MESSING R B; LYTLE L D
US 4594358 A US 4035511 A 77-52098Y/29 (MASI ) MASSACHUSETTS
INST TECHNOLOGY
US 4594358 A US 4083982 A 78-40255A/22 (MASI ) MASSACHUSETTS
INST TECHNOLOGY/MESSING R B; LYTLE L D
US 4594358 A US 4313896 A 82-13654E/07 (ELIL ) LILLY & CO ELI/
MOLLOY B B; SCHMIEGEL K K
US 4594358 A US 4314081 A 75-49665W/30 (ELIL ) LILLY & CO ELI
CITED LITERATURE REFERENCES
Family Member Cat Citation
By Examiner:
US 4594358 A Merck Index, 9th Ed (1976) pp. 1015-1016.
US 4594358 A Hynes et al., Drug Development Research, 2, 33
(1982).
US 4594358 A Messing et al., Psychopharmacology Communications,
1(5), 511 (1975).
US 4594358 A Larson et al., Life Sciences, 21, 1807 (1977).
US 4594358 A Sugrue et al., J. Pharm. Pharmac., 28, 447 (1976).
6/2/18
DIALOG(R)File 342:Derwent Patents Citation Indx
(c) 2001 Derwent Info Ltd. All rts. reserv.
00186200 WPI Acc No: 84-252258/41
Use of fluoxetine or norfluoxetine as anti-anxiety agents - pref.
administered orally as hydrochloride salts
Patent Assignee: (ELIL ) LILLY & CO ELI
Author (Inventor): STARK P
Patent Family:
Patent No Kind Date Examiner Field of Search
GB 2137496 A 841010 (BASIC)
AU 8426457 A 841011
DE 3413093 A 841011
DE 3467704 G 880107
DK 166479 B 930601
DK 8401132 A 841009
EP 123469 A 841031
EP 123469 B 871125
GB 2137496 B 861001
IT 1175977 B 870812
JP 59193821 A 841102
JP 94035382 B2 940511
US 4590213 A 860520
ZA 8402457 A 851002
Derwent Week (Basic): 8441
Priority Data: US 483087 (830408)
Applications: US 483087 (830408); DK 841132 (840228); ZA 842457 (840402);
JP 8467310 (840403); DE 3413093 (840406); EP 84302361 (840406); GB
848880 (840406); GB 84408880 (840406)
Designated States
(Regional): BE; CH; DE; FR; GB; IT; LI; LU; NL; SE
Derwent Class: B05
Int Pat Class: A61K-031/13; A61K-031/135
Number of Patents: 014
Number of Countries: 015
Number of Cited Patents: 009
Number of Cited Literature References: 000
Number of Citing Patents: 022
CITED PATENTS
Family Member Cited Patent Cat WPI Acc No Assignee/Inventor
By Examiner:
EP 123469 A US 4194009 A 75-49665W/30 (ELIL ) LILLY & CO ELI
EP 123469 A US 4313896 A 82-13654E/07 (ELIL ) LILLY & CO ELI/
MOLLOY B B; SCHMIEGEL K K
EP 123469 B US 4194009 A 75-49665W/30 (ELIL ) LILLY & CO ELI
EP 123469 B US 4313896 A 82-13654E/07 (ELIL ) LILLY & CO ELI/
MOLLOY B B; SCHMIEGEL K K
US 4590213 A US 4018895 A 75-49665W/30 (ELIL ) LILLY & CO ELI
US 4590213 A US 4194009 A 75-49665W/30 (ELIL ) LILLY & CO ELI
US 4590213 A US 4313896 A 82-13654E/07 (ELIL ) LILLY & CO ELI/
MOLLOY B B; SCHMIEGEL K K
US 4590213 A US 4314081 A 75-49665W/30 (ELIL ) LILLY & CO ELI
US 4590213 A US 4329356 A 82-43710E/21 (ELIL ) LILLY & CO ELI/
HOLLAND D R
?