“Walking pathways” and how promoters can help to find new drugs .

1. “Walking pathways” and how promoters can help to find new drugs. (Practical guide to multi-omics and multi-scale data integration) Alexander Kel Biosoft.ru, Skolkovo Moscow Wolfenbüttel Novosibirsk alexander.kel@genexplain.com

2. Trovafloxacin - antibiotic Withdrawn from market due to risk of idiosyncratichepatotoxicity in 2001.

3. Failure Affects National Economies:Medicines & Equitable Distribution Total: 1.3 T$ Combined treatment and productivity costs for US in 2007 Milken Institute 2008 Hensley

5. One of the main causes of high death rate for such diseases is the unsatisfactory quality of treatment, which in the first place is brought by low efficiency and insufficient safety of today’s drugs and therapies. About 50% of prescribed medicine doesn’t have any therapeutic effect at all. Moreover, 125 thousand deaths annually (in USA) are caused by the drugs’ side effects. It becomes more and more obvious that the main cause of this crisis is the insufficient understanding of deep biological mechanisms of initiation and flowing of pathological conditions and toxicity mechanisms used in drugs.

6. Drug discovery – the Gold Rush

7. Drug discovery – shouldbecome a technology Disease Therapy Patient

8. Systems medicine Systems approaches will transform the way drugs are developed … that will target multiple components of networks and pathways perturbed in diseases. They will enable medicine to become predictive, personalized, preventive and participatory Systems medicine: the future of medical genomics and healthcare Charles Auffray1*, Zhu Chen2 and Leroy Hood3 Genome Med 2009, 1:2

9. Weshouldfind a keypathwayof a disease, selecta goodtargetandinhibit it. TRANSPATH

10. Pathway mapping Mapping on pathways Cause of disease ?? Differentially expressed genes/proteins

11. TNF-a 117 differentially expressed genes

12. Can we predict TNF pathway? ? 117 differentially expressed genes

13. Canonical TNF pathway TRANSPATH

14. Lets do mappingthedifferentiallyexpressed genes on canonicalpathways. Not significant Not significant TNF pathwaycan not befoundbydirectmaping on canonicalpathways....

15. Human epidermoid carcinoma A431 cells treated by epidermal growth factor (EGF) 320 differntially expressed proteins EGF

16. Mapping differentially expressed proteins to canonical signal transduction pathways

17. Mapping on pathwaysdoes not work(even in such a simple cases) Why ?

18. Pathwaysarefarfrombeing fullyundersood.

19. BIG gap of knowledge on interactions between TF and their target sites in DNA TF2 TF3 TF1 TRANSPATH

20. ? Search for new TF binding sites with PWMs (Match algorithm) …

21. N M k s n=k+s p=M/N

22. Overrepresented TFs in TNF-alpha regulated promoters

24. Master regulator

25. Search for the reason by the analysis of the ripples

26. Can we predict TNF pathway? ? 117 differntially expressed genes

27. GeneXplain platform – drug target discovery pipeline

28. TNF-alpha

29. Human epidermoid carcinoma A431 cells treated by epidermal growth factor (EGF) ? 320 differntially expressed proteins EGF Master regulatoranalysis EGF was still not in thelist !

30. Pathwaysarefar .....far....farfrombeingfully undersood!

31. Combinatorial regulation „Fuzzy puzzle“

32. AP-1 ******* TGAGTCA Human collagenase (-2013) ** ** * TGTGTAA Mouse IL-2 (-143) ** * TGTAATA Mouse IL-2 (-82) TGAgTCA Consensus:

33. Mouse c-fos promoter (Matrix search for TF binding sites)

34. Mouse Interleukin-2 gene promoter AP-1 COMPEL:C00050 NF-ATp . . . . . . . tgccacacaggtagactcttTTGAAAATAtgTGTAATAtgtaaaa catcgtgaca cccccatatt… … -96 -79 ST TGAGTCA One of the TF binding sites in a composite elements can be rather weak.Weak DNA-protein interactions are stabilized by protein-protein interactions. AP-1 consensus

35. Composite Module (CM) Composite Modules (CM) (Mark Ptashne, Alexander Gann Genes and Signals, 2002)

36. Composite Modules (CM) w Start of transcription ... ( ) ( ) k k s s ... 1 mk ... ... ... Parameters of the model to be estimated by GA ... Wecreated a geneticalgorithmto find sitecombinations

37. Composite Modules (CM) Composite Module Score (cms) K, the number of individual PWMs in the module, (k=1,K) ... Matrix cut-off values: Relative impact values: Maximal number of best matches: mk R, the number of pairs of PWMs (r=1,R) Matrix cut-off values: Relative impact values: Maximal and minimal distances:

38. Fitness function of the Genetic-Regression Algorithm (GRA) # promoters R – linear regression N FN – false negatives FP – false positives FN cms FP T – T-test (differencebetweenmeanvalues) N – normal likeness k – number of free parameters

39. Composite module in promoters of cell cycle-related genes Cell cycle-related promoters Background sequences

40. Mouse c-fos promoter

41. Promoter structure: curret paradigm Mouse IL-4 promoter AP-1 HMG Y c-MAF AP-1 HMG Y NF-Y STAT 6 AP-1 AP-1 TATA TSS NFAT NFAT NFAT NFAT NFAT CE CE -249 -114 -180 -150 -88 -60 -28 +1

42. Promoter is a parking place

43. Promoter structure: reality Mouse c-fos promoter

44. Parking in Italy

45. ChIP-seq data: EPS-FLI1 Ewing sarcoma transcription factor – gene fusion

47. Network plasticity „Walking pathways“