280 likes | 303 Views
Chemoinformatics and Metabolism. Paula de Matos. Chemoinformatics and Metabolism Group Research. Indexing, searching and dissemination of chemical information. Natural Products and Metabolomics. Cheminformatics Algorithms and Toolkits. Chemical Entities of Biological Interest.
E N D
Chemoinformatics and Metabolism Paula de Matos
Chemoinformatics and Metabolism Group Research Indexing, searching and dissemination of chemical information Natural Products and Metabolomics Cheminformatics Algorithms and Toolkits
Chemical Entities of Biological Interest • A database containing a freely available, manually annotated dictionary of molecular entities focused on ‘small’ chemical compounds. • Provides a method to navigate the chemical space via an ontology • ChEBI aims to provide a central, definitive reference of chemical nomenclature. Dictionary
http://www.ebi.ac.uk/chebi Dictionary Resource for Nomenclature
What does ChEBI cover? • Mostly small entities • Big entities too like • alumina • amylose • metaborate • Excludes proteins and nucleic acids 5
ChEBI Web Services • Programmatic access to a ChEBI entry • SOAP based Java implementation • Clients currently available in Java and perl • Four methods with which to access data • getLiteEntity • getCompleteEntity • getOntologyParents • getOntologyChildren • Documented at http://www.ebi.ac.uk/chebi/webServices.do.
ChEBI further info • http://www.ebi.ac.uk/chebi • Mailing lists: • chebi-help@ebi.ac.uk • chebi-announce@lists.sourceforge.net • chebi-ontology@lists.sourceforge.net • Submitting data • http://www.ebi.ac.uk/chebi/submissions
The Chemistry Development Kit (CDK): An Open Source Java-Library for Structural Chemo- and Bioinformatics • >90.000 Lines of Code, >900 Classes, > 9000 Methods • Library Generation • Virtual Screening • Molecular Property Prediction • Visualization http://cdk.sourceforge.net (1) Steinbeck, C.; Hoppe, C.; Kuhn, S.; Guha, R.; Willighagen, E. L. Current Pharmaceutical Design 2006, 12, 2111-2120. (2) Steinbeck<, C.; Han, Y. Q.; Kuhn, S.; Horlacher, O.; Luttmann, E., Willighagen, E. Journal of Chemical Information and Computer Sciences 2003, 43, 493-500.
The Chemistry Development Kit (CDK) Visualization Input/Output • I/O (CML, MDL Molfile, SDF, PDB) • SMILES • InChI • Structure-Diagram-Layout (SDG) • 2D Rendering • 3D Rendering Library Enumeration Modelling • Deterministic Isomer generator • Stochastic Structure Generators via • Simulated Annealing • Genetic Algorithms • 3D Model-Builder • Atom-Typing • Force-Field • Representation of Biomolecular Structures Chemical Graphs Properties • Isomorphism detection • Maximum-Common-Substructure Searches • SMARTS- and Substructure searches • Ring searches • Aromaticity detection • Fingerprinting • > 70 QSAR-Descriptors • QSAR model building
Example: Fingerprinting IMolecule superstructure = MoleculeFactory.makeIndole(); IMolecule substructure = MoleculeFactory.makePyrrole(); Fingerprinter fingerprinter = new Fingerprinter(); BitSet superBS = fingerprinter.getFingerprint(superstructure); BitSet subBS = fingerprinter.getFingerprint(substructure); boolean isSubset = FingerprinterTool.isSubset(superBS, subBS); Bitscreen coding for structural features 0 0 1 1 0 1 0 0 1 0 Hetero-aryl -COOH Alky O-Alkyl -NH2
CDK in numbers • 67 registered developers on SF • 86 people subscribed to cdk-devel list • 111 people subscribed to cdk-user list
CDK in numbers 80,966 downloads since 2001
CDK in numbers CDK article (2003) cited 68 times
CDK info • Project home page: • http://cdk.sourceforge.net/ • Mailing list: • cdk-user@lists.sourceforge.net • cdk-devel@lists.sourceforge.net • Documentation • http://pele.farmbio.uu.se/nightly/
OrChem • Oracle chemistry plug-in using the Chemistry Development Kit (CDK) providing substructure and similarity searches for chemical graphs. • OrChem is suitable for Oracle 11G and onwards • Not an Oracle data cartridge - it doesn't need Oracle's extensibility architecture because its Java components run as Java stored procedures inside the Oracle standard JVM (Aurora).
Problem • Chemical substructure or similarity searching is computationally expensive especially on a large dataset?
Example OrChem Queries • Similarity search • select * from table( orchem_simsearch.search( 'OC4=C(C(=C3OC(C)(COC=1C=CC(=CC=1)CC2C(=O)NC(=O)S2)CCC3=C4C)C)C','SMILES',0.8,null,'N') ) ; • Substructure search • select orchem_subsearch.search(molfile,'MOL',50,'Y') from compounds where molregno=12345;
Parallel vs. Non parallel Performance of substructure search on 3.5 million compounds
Substructure benchmarking Performance of substructure search on 3.5 million compounds
OrChem info • http://orchem.sourceforge.net/ • Mailing list: • orchem-devel@lists.sourceforge.net