1 / 29

Compound Library Annotation

Compound Library Annotation. Miklós Vargyas. May , 200 5. Compound Library Annotation. Overview of the Screen Package Virtual screening Optimized dissimilarity metrics Clustering Library Annotation – a real-life application of the Screen tool Approach #1: use command line applications

albina
Download Presentation

Compound Library Annotation

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. Compound Library Annotation Miklós Vargyas May, 2005

  2. Compound Library Annotation • Overview of the Screen Package • Virtual screening • Optimized dissimilarity metrics • Clustering • Library Annotation – a real-life application of the Screen tool • Approach #1: • use command line applications • Approach #2: • API programming

  3. Overview of the Screen Package 0100010100011101010000110000101000010011000010100000000100100000 0001101110011101111110100000100010000110110110000000100110100000 0100010100110100010000000010000000010010000000100100001000101000 0101110100110101010111111000010000011111100010000100001000101000 0001000100010100010100100000000000001010000010000100000100000000 0100010100010100000000000000101000010010000000000100000000000000 0101010101111100111110100000000000011010100011100100001100101000 0100010100011000010000011000000000010001000000110000000001100000 0000000100000000010000100000000000001010100000000100000100100000 0101110100110101010111111000010000011111100010000100001000101000 queries hypothesis fingerprint metric 0000000100001101000000101010000000000110000010000100001000001000 0100010110010010010110011010011100111101000000110000000110001000 0100010100011101010000110000101000010011000010100000000100100000 0001101110011101111110100000100010000110110110000000100110100000 0100010100110100010000000010000000010010000000100100001000101000 0100011100011101000100001011101100110110010010001101001100001000 0101110100110101010111111000010000011111100010000100001000101000 0100010100111101010000100010000000010010000010100100001000101000 0001000100010100010100100000000000001010000010000100000100000000 0100010100010011000000000000000000010100000010000000000000000000 0100010100010100000000000000101000010010000000000100000000000000 0101010101111100111110100000000000011010100011100100001100101000 0100010100011000010000011000000000010001000000110000000001100000 0000000100000000010000100000000000001010100000000100000100100000 0100010100010100000000100000000000010000000000000100001000011000 0001000100001100010010100000010100101011100010000100001000101000 0100011100010100010000100001001110010010000010001100000000101000 0101010100010100010100100000000000010010000010010100100100010000 Virtual hits targets targetfingerprints

  4. Need for Optimization 0.57 0.55 0.47

  5. Optimized Metrics asymmetry factor scaling factor asymmetry factor weights

  6. Improved Similarity by Optimization 0.57 0.55 0.47 0.20 0.06 0.28

  7. Enrichment by Optimization

  8. Clustering 8 active compound sets • ACE inhibitors • angiotensin 2 antagonists • D2 antagonists • delta antagonists • FTP antagonists • mGluR1 antagonists • Thrombin inhibitors • 5-HT3-antagonists

  9. Ward Centroids

  10. Maximum Common Substructure Clustering

  11. Compound Library Annotation Annotate library: predicted activity in some therapeutic areas ActACE=0.5 Actß2=0.98 Actß2=0.45 ActACE=0.78

  12. Similarity Based Activity Prediction Use sets of known actives to predict activity of compounds ActACE=0.55 0101110100110101010111111000010000011111100010000100001000101000 Actß2=0.98 0101110100110101010111111000010000011111100010000100001000101000

  13. Approach #1: Off the Shelf ChemAxon Tools • Parameter setting • Pharmacophore fingerprint • Tanimoto dissimilarity metric • Median Pharmacophore Hypothesis screenmd library.sdf ace.sdf \ –o SDF annotated-library.sdf \ -k PF –M Tanimoto –H Median

  14. Single Active Family

  15. Multiple Active Families screenmd library.sdf ace.sdf \ -o SDF lib-ace.sdf -k PF –M Tanimoto –H Median screenmd lib-ace.sdf beta2.sdf \ -o SDF lib-ace+beta2.sdf -k PF –M Tanimoto –H Median screenmd lib-ace+beta2.sdf delta.sdf \ -o SDF lib-ace+beta2+delta.sdf -k PF –M Tanimoto \ –H Median screenmd lib-ace+beta2+delta.sdf D2.sdf \ -o SDF lib-ace+beta2+delta+D2.sdf -k PF –M Tanimoto \ –H Median ... ...

  16. Annotated Library File

  17. Approach #2: Using ChemAxon JChem API • API programming – custom solution • PharmacophoreFingerprint and the MolecularDescriptor class hierarchy • Tanimoto dissimilarity calculation • Median Hypothesis calculation • Description generation for structure in SDfile • Writing structures in SDfile

  18. MolecularDescriptor class hierarchy Molecular Descriptor Pharmacophore Fingerprint BCUT Chemical Fingerprint CUSTOM MACCS

  19. MolecularDescriptor Sets Molecular Descriptor Set Molecular Descriptor 1 (e.g. CFp) Molecular Descriptor 2 (e.g. PFp) Molecular Descriptor 3 (e.g. logP)

  20. Dissimilarity Calculation MDSet s1 = MDSet.newInstance( new String[]{“CF”,”PF”,”LogP”} ) MDSet s2 = MDSet.newInstance( new String[]{“CF”,”PF”,”LogP”} ) . . . Generate s1 and s2 somehow . . . System.out.println( “dissimilarity(s1,s2) = “ + s1.getDissimilarity( s2 ) );

  21. Tanimoto Dissimilarity Calculation MDSet s1 = MDSet.newInstance( new String[]{“PF”} ) MDSet s2 = MDSet.newInstance( new String[]{“PF“} ) . . . Generate s1 and s2 somehow . . . PharmacophoreFingerprint pf1 = s1.getDescriptor(0); PharmacophoreFingerprint pf2 = s2.getDescriptor(0); System.out.println( “Tanimoto(pf1,pf2) = “ + pf1.getTanimoto( pf2 ) );

  22. Median Hypothesis Calculation MDSet s1 = MDSet.newInstance( new String[]{“PF”} ) MDSet s2 = MDSet.newInstance( new String[]{“PF“} ) . . . Generate s1 and s2 somehow . . . MDHypothesisGenerator medianHypoGenerator = MDHypothesisCreator.create( "Median" ); medianHypoGenerator.add( s1 ); medianHypoGenerator.add( s2 ); MDSet hypothesis = medianHypoGenerator.generate();

  23. Reading Descriptors from Structure File MDFileReader inputReader = new MDFileReader( “library.sdf”, MDSet.newInstance( new String[]{"PF"} ) ); MDSet mdRead = inputReader.next();

  24. Writing structures in SDfile MolExporter outputWriter = new MolExporter( new PrintStream( new BufferedOutputStream( new FileOutputStream( fileName ))), "sdf"); Molecule m = getAMolecule(); outputWriter.write( m );

  25. LibAnnot class Full source code avaialable at http://www.chemaxon.com/

  26. Future plans • New MolecularDescriptors (e.g. 3D Pharmacophore) • Non-hierarchical MCS clustering, better GUI • Library diversity estimation

  27. Summary • Screen+JKlustor for optimized virtual screening and hit set profiling • Library annotation by screenmd • Library annotation by custom program

  28. Acknowledgements and Credits • JKlustor developed by Ferenc Csizmadia et al • Optimizer developed by Zsuzsa Szabó • PMapper developed by Szilárd Dóránt, Nóra Máté • Pharmacophore definitions by György Pirok

  29. Thank you for your attention Máramaros köz 3/a Budapest, 1037Hungaryinfo@chemaxon.comwww.chemaxon.com

More Related