C371 Chemical Informatics I Projects and Papers

Updated: 8 November 2004

The following projects have been selected:

Elizabeth Adams: de novo Drug Design paper
Philip Bidwell: Protein Folding paper
Matt Carrico: Proteomics paper
Patrick Coffman: application of 3D QSAR (CoMFA) for the design of new drugs paper
Tasia Pyburn: de novo Drug Design paper
Manuel Rodriguez: Chemistry Development Kit (CDK)
Matt Spraker: Chem TK Lite
Daniel Vu: Project work with Dr. Mookie Baik

• Creation of a reference database for Beilstein Lawson Numbers.
  The Beilstein system is based on the constitutional formula of the compound as only classifying feature. It takes into account only the structural properties (the morphology) of the compound. Thus, the Beilstein system represents a one-dimensional, natural, hierarchical classification.

  The principal criteria for classifying compounds in the system are:

  1. main division, i.e., acyclic, iso-(carbo-)cyclic, and heterocyclic (further subdivided according to the type and number of ring heteroatoms)
  • Acyclic Compounds, Volumes 1-4; System Numbers 1-449
  • Isocyclic Compounds, Volumes 5-16; System Numbers 450-2358
  • Heterocyclic Compounds, Volumes 17-27; System Numbers 2359-4720.
  2. functional groups, i.e., no functional group, OH, C=O, COOH, NH2, etc.
  3. degree of saturation (CnH2n-x)
  4. C number (total number of carbon atoms)
  5. skeletal structure (unbranched, branched, etc.; monocyclic, bicyclic, etc.)

  Advantages of using the system to classify organic compounds are:

  • Every compound (i.e., every structural formula) is assigned a unique place in the Beilstein handbook.
  • Every compound can be retrieved from the Beilstein handbook without use of indexes or nomenclature.
  • Chemically related compounds are brought into close proximity by the system, thus enabling the Beilstein user to quicky find "chemical relatives" (homologs, analogs) of every compound searched for, a kind of "similarity search".

  In the Beilstein database the LAWSON NUMBER is:

  • based on the Beilstein System
  • used for similarity searches
  • represents structure fragments.

  Every substance in the Beilstein file has at least one Lawson number, but searching by Lawson Numbers is NOT substructure or Markush searching. However, the Lawson Number does represent certain structural fragments and can thus be used for structural similarity searches. In general, the smaller the Lawson Number, the more common the fragment. Dividing the Lawson Number by 8 puts you roughly in the Beilstein system number for the printed Beilstein volume that contains the compound.

  Unfortunately, the Beilstein Institute never published the meanings of the 4,720 system numbers used to classify organic compounds. However, the Lawson Number could serve as an effective index search key for the compounds if its meaning were known. Searches that include the Lawson Number are effective when used in combination with other search keys, such as molecular formula, element ranges, etc. It is also useful when combined with NOT in substructure searches.

  This project will refine and improve a database of Lawson Numbers created by a student last year to link LNs to their meanings. In its present form, the database can be seen at:

  http://mypage.iu.edu/~ucoca/begperl/formFetch.html

  These are some of the known problems with the database, as it presently exists:

  • Some strings do not work, such as: As,Bi(*n)/P[V]/P[III]/N=N-
  • A second problem is the truncation of both the display and search of all multi-word entries. Examples are:
    • Parent Carbocyclics
    • Parent Heterocycles O(3)
    • Parent Heterocycles O(4)
  • The search appears to be case sensitive. For example, N-R/O-R/(O1) retrieves entries, but n-r/o-r/(o1) does not.
  If you choose to work on this project, we will provide you with the raw data.
• Investigation of ChemTK for chemical screening data
  ChemTK Lite is a free Windows toolkit providing chemists the ability to visualize, organize, query, and analyze chemical screening data. We use ChemTK Lite on one of the exercises in C371. We expect to have the full version of ChemTK available for projects. If you choose this topic, you will have to learn how to use ChemTK and explain it to the class.

• Investigation of the INChI Code for chemical substances
  The source code and Application Program Interface (API) for the test version of the IUPAC-NIST Chemical Identifier (INChI) are now available. The source code is written in ANSI C. It may be used in a DLL (dynamic link library) or as a directly linkable library. It is accompanied by examples of calling programs written in ANSI C and Microsoft Visual Basic 6.

  The source code and examples of calling programs are available for download at:

  http://chemdata.nist.gov/IChI/INChI_v112B_API.zip

• Investigation of the Chemistry Development Kit (CDK), an open-source Java library for chemo- and bioinformatics
  The CDK is described in an article by Christoph Steinbeck and others in the Journal of Chemical Information and Computer Sciences, 2003, 43, 493-500. Demonstrate how the CDK implements tasks such as 2D and 3D rendering of chemical structures, SMILES parsing and generation, ring searches, isomorphism checking, structure diagram generation, etc.
• Investigation of OpenEye software
  The OEChem software is loaded in Informatics at IUB on Xavier in the publicly accessible directory /usr/local/oechem/. There are C++ and Python libraries, example code, and documentation there.
• Investigation of the appropriateness of the Expereact Web program to maintain a local database of chemical products that can be shared over the Internet
  http://www.chemexper.com/xtweb/main.shtml is the site where the free copy of Expereact WEB can be obtained. How does it relate to MolBank and the Budapest Open Access Initiative for publishing journals?
• OpenBabel vs. CLiDE vs. VEGA as tools for converting between chemical file formats
  Babel and CLiDE have features to recognize and/or interconvert common file formats for coding chemical structures. For information on OpenBabel, see http://openbabel.sourceforge.net. See Henry Rzepa's August 15, 2003 posting on chemweb for compiled versions of OpenBabel for MacOS X, SGI (Irix 6.5), Linux (Redhat 8), and Windows versions. http://mailman.ic.ac.uk/pipermail/chemweb/ Information on the CLiDE program, available for Windows, can be found at http://www.simbiosys.ca. Information on VEGA is found at http://users.unimi.it/~ddl
• RSS applications to Chemistry
  See "Towards the Chemical Semantic Web: An Introduction to RSS," by Peter Murray-Rust and Henry Rzepa at: http://www.ch.ic.ac.uk/rzepa/rss/. Also, Rzepa's postings of April 29, June 13, and June 27, 2003 and later on the Chemweb archives: http://mailman.ic.ac.uk/pipermail/chemweb/

  Crystallography Journals Online now publishes RSS 1.0 feeds for certain IUCr journals: http://www.iucr.ac.uk/services/rss.html

• Comparison of the ChemWeb.com Compound Search at http://www.chemweb.com/compounds and NLM's structure searching site at: http://chem.sis.nlm.nih.gov/chemidplus/setupenv.html
• Comparison of Deep View (Swiss PDBviewer) and RasMol
  Deep View has been said to be both easier and more powerful than Rasmol, and it is supposed to talk naturally to many other bioinformatics tools, like structure and sequence databases. See a Deep View tutorial at:

  http://www.usm.maine.edu/~rhodes/SPVTut/index.html

  On the other hand, Rasmol http://www.openrasmol.org/ has been around a long time and formed the basis for MDL's widely used Chime program.

• Evaluation of chemistry software for Linux
  One source of information is:
  http://zeus.polsl.gliwice.pl/~nikodem/linux4chemistry.html
  



• A Cheap Chemical Information System

  The basic assumption in this project is that we have little or no money to work with. Hence, we need to build a prototype information system that has free or nearly free components available from the Web. Given that constraint, it may not be possible to include all desirable features in the prototype system. However, it will at least include information (or how to gain access to information) or software to assist with:

  • Chemical nomenclature
  • Chemical drawing programs
  • Personal database software (including bibliography managers)
  • Data analysis software
  • Electronic laboratory notebooks and LIMS (laboratory information management systems)
  • Chemical storeroom management
  • Identification of the commercial products that include molecular modeling tools as part of an integrated chemical information system. (Some companies to consider: Accelrys, Bio-Rad, CambridgeSoft, Chemical Design, Cherwell, Daylight, Fujitsu, Gaussian, Hypercube, Interactive Simulations, MDL, Serena Software, Q-Chem, Semichem, Tripos, and Wavefunction.)

  Features should include:

  • Ease of use
  • Ability to input a chemical structure as a search key and search an appropriate database
  • One component should be a guide to useful free and commercial chemistry software covering the areas outlined above.

  This project requires you to look at what is available commercially and for free in the chemical software area and design a system that will let a chemist find the appropriate tools to take an idea from conception to publication. That means providing access to free software, such as ISIS/Draw, to create chemical structures for publication, to databases or software that would help the chemist identify appropriate reagents and suppliers of those reagents, allow the creation of a chemical database to manage all of the information gathered, and to keep track of progress as the project moves along.

  Your first task is to discover what is available and to begin to categorize the types of tools. You will have to connect to the Web sites of the companies and do a lot of investigation.

  Start by reading some of the lectures in the C471 Chemical Information Sources and Services course at: http://www.indiana.edu/~cheminfo/C471/471sched.html, especially the lesson on Science Writing Aids and the structure drawing tools that are used in that lesson.

  The SIRCh page on Molecular Visualization Tools and Sites at: http://www.indiana.edu/~cheminfo/ca_mvts.html should give you some additional ideas for tools to incorporate into your chemical information system.

  MDL's (http://www.mdl.com) Report Manager product has some of the features that should be in your system. For a data model, you might want to consider NCI's DTP Data Warehouse Model.

PAPERS

• Comparison of Standard file formats for numeric, spectral, and crystallographic data.

  JCAMP-DX for spectral data: http://www.irug.org/documents/JCAMP_IRUGweb.pdf
  http://wwwchem.uwimona.edu.jm:1104/software/jcampdx.html

  CIF for crystallographic data: http://www.iucr.ac.uk/iucr-top/index.html

  PDB File Format: http://www.rcsb.org/pdb/info.html#File_Formats_and_Standards

• Protein Folding (review article)
  What is the current state of the art in protein folding? (Broad or deep coverage, but not both)

• De novo Drug Design (review article)
  What specific progress can we claim with de novo drug design at this point in time?

• Bioinformatics Software and Databases from Commercial Companies
  What aspects of commercial software (i.e., not free and publicly available on the Web) offer significant enhancements to the features that can be had from the free options?

• Describe the applications of graph theory to the processing of chemical structure databases.
• What factors must be taken into account when designing a screening system for chemical substructure searching.
• Discuss the application of similarity searching in the discovery of novel bioactive compounds.
• Describe the services and systems currently available for the retrieval of information from chemical patents.
• Discuss the application of either classical QSAR (Hansch analysis) or 3D QSAR (CoMFA) to the design of new drugs.
• Describe the mode of operation of either: docking programs; or de novo ligand design programs.
• Discuss the application of compound selection strategies and molecular diversity in combinatorial chemistry.
• In relation to nucleic acid and protein sequence databases, detail the theory and application of sequence alignment.
• Give an overview of the significance of database interoperation to biologists. What practical difficulties are involved and how are these being addressed?

PAST PROJECTS

In past years, C371 students have worked on the following projects:

I. Digitizing and Indexing Groth's Chemische Krystallographie.

This old five-volume German set is in the IUB Chemistry Library Reference Area with call number QD951.G8. It has recently been found to be a useful resource by organic chemists who are investigating inclusion compounds. Since the work was published prior to 1923, there is no copyright restriction on scanning and placing it on the Web.

Participants must choose an appropriate format and scan the material in such a manner as to allow for easy retrieval on the Web. Consideration must be given to file size and download times, as well as indexing.

The work itself, of course, is in German, so one of our tasks will be to determine English IUPAC names, CAS Registry Numbers, and other appropriate chemical identifiers for the substances. SMILES codes are to be constructed for each.

This is a long-term project that will likely stretch over more than one C371 class, so we must make judicious choices on the approach to the creation of this digital library item. Thus, consultation with people in and use of the resources developed by the IU Libraries/UITS Digital Library Program is advisable.

II. Reciprocal Net and the Common Molecules Project

The IUB Molecular Structure Center has for several years been developing a project known as Reciprocal Net. In particular, the Common Molecules portion of the project is one that could use the assistance of C371 students.

III. Importing Chemical Data Into Excel or Bibliography Manager Software

Large databases such as Beilstein or Chemical Abstracts have prodigious amounts of data and many options for retrieval of the data. However, chemists often want to massage the data and/or reuse them for other purposes. In this project, the students will learn to use effectively a standard spreadsheet program (Microsoft Excel) and one of three general software packages (EndNote, ProCite, or Reference Manager) to import data from commercial chemical databases and the Web.

Some of the major database vendors have developed ways of effectively downloading data for importing to other packages, but those do not always work well for chemical data. Your task will be to document the procedures for retrieving and importing the data and to see which, if any, of the standard software packages allow the importing of structure data that can subsequently be displayed as chemical structures.

IV. A Prototype System for Reformatting CA File Data

Past SLIS students have attempted to extract data from a search of Chemical Abstracts and run it through a Perl script that in effect punctuates the records into The ACS Style Guide format.