Transcript Drug - CBS
ChemProt: A Disease Chemical Biology Database Sonny Kim Nielsen – PhD student Computational Chemical Biology - CBS Department of Systems Biology - DTU [email protected] Translational Informatics, january 2011 PLAN Drug-Target interactions and Target-Disease associations Chemical Biology repository (ChemProt) Biological Network and visualisation using Cytoscape We hope for a simple concept… Drug Gene Phenotype But in reality it is not so simple Phenotype Phenotype Gene Gene Gene Phenotype Gene Phenotype Drug Gene Gene Phenotype Gene Gene Phenotype Phenotype What is the number of targets for a drug? 4400 drugs, 2.7 targets/drug in average 1081 drugs, 5.69 targets/ drug in average Wombat-PK The pharmacology of a drug is still sparse Chemical similarity Compounds Proteins Garcia-Serna R et al. Bioinformatics 2010 Keiser MJ et al. Nat. Biotech 2007 What about phenotypes? Protein-Protein interactions Network (PPI) disease Especially genetic disorders (color blindness, Huntington’s disease, Cystic fibrosis) disease Prader-Willi syndrome (7 genes) disease Cancer, diabetes, mental illness A quality-controlled human protein interaction network Download and reformat PPI databases Trans organism ppi transferral Automated scoring of all interactions 500 000 interactions between 10,300 human proteins Lage et al. Nat. Biotech 2007 Tissue-specific pathology and gene expression of human disease genes and complexes Lage K, et al.,PNAS, 52, 20870-20875 (2008) Integration of chemical space into biological space Chemical space Phenomeinteractome Chemical bioactivity Disease chemical biology network ChemProt: a disease chemical biology database 700.000 unique bioactive compounds for 30.000 proteins MACCs and pharmacophore fingerprints computed Protein-Protein Interactions data integration** Structural similarity search 428000 PPIs Disease-protein complexes Compound 1 Compound 2 D1 D2 P1 P2 Protein 1 Protein 2 OMIM, AKS2, GO, Tissue specificity Taboureau O et al. 2011 Nucleic Acids Res. An example with citalopram (antidepressant) Drug Bank Chembl 7 proteins ChemProt PPIs disease 49 proteins 629 genes forming 4141 interactions An example with citalopram: Genes enrichment GO OMIM AKS2 - cell communication (p-value 1.32 e-86) - signal transduction (p-value 4.07 e-81) - Major Depressive Disorder (p-value 3.77 e-06) TPH2;FKBP5;HTR2A - Obsessive-Compulsive Disorder 1 (p-value 1.67 e-04) HTR2A;SLC6A4 - Schizophrenia 9.02 e-24 - Bipolar disorder 6.92 e-21 - Anorexia nervosa 6.61 e-10 - Bulimia nervosa 1.41 e-07 - Obesity 2.20 e-05 An example with citalopram: Genes enrichment for DRD4 Leukemia An example with citalopram: Src regulation of hERG? Citalopram inhibits hERG (25 μM in ChEMBL) Associated to LQTS and arrhythmia Future plan in ChemProt: Drugs categorization based on the Anatomical Therapeutic Chemical (ATC) Classification. In ATC classification system, the active substances are divided into different group according to the organ or system on which they act and their therapeutic, pharmacologcal and chemical properties Drug – Target – Disease Classification through ATC Drugs categorization based on the Anatomical Therapeutic Classification (ATC) Disease – Disease Network Disease – Disease Network Conclusion Systems chemical biology allow to investigate new direction in understanding molecular (adverse) effects of chemicals in biological systems. Some technical information about biological network and Cytoscape. Biological networks in Bioinformatics Graph theory Graph G=(V, E) is a set of vertices V (nodes) and edges E A subgraph G´of G is induced by some V´ V and E´ E Graph properties: - neighborhood - Connectivity (node degree, paths) - Directed vs. undirected Path theory A path is a sequence {X1, X2,…, Xn} such that (X1, X2), (X2, X3),…,(Xn-1, Xn) are edges of the graph. A closed path Xn= X1 on a graph is called a graph cycle or circuit. Biological networks in Bioinformatics Protein network representations Biological networks in Bioinformatics Sparse vs dense Clustering coefficient How to represent interaction network? Network visualization and analysis tool. A community based framework for networks modeling www.cytoscape.org Cytoscape Desktop Network Management panel Network overview panel Network graph and view Attribute browser Input – output data Cytoscape reads an interaction network: – Using a interaction file (.sif, .txt, .csv…) – A XML format (essentially to store information and allows network data exchange with a variety of other network display programs A A B A D B B D C – Output can be a .sif format, txt, but also image (png, pdf, jpeg…) C C Input – output data Add information on nodes and edges Visual style (Vizmapper) - Layout Time for you to play And don’t stress…