Jack Tuszynski Accelerating Chemotherapy Drug Discovery with Analytics and High Performance Computing
•Télécharger en tant que PPT, PDF•
1 j'aime•1,205 vues
Jack Tuszynski's Best of Analytics presentation May 14, 2013 "Accelerating Chemotherapy Drug Discovery with Analytics and High Performance Computing."
Recommandé
Recommandé
Contenu connexe
Tendances
Tendances (20)
Similaire à Jack Tuszynski Accelerating Chemotherapy Drug Discovery with Analytics and High Performance Computing
Similaire à Jack Tuszynski Accelerating Chemotherapy Drug Discovery with Analytics and High Performance Computing (20)
Plus de Kim Solez ,
Plus de Kim Solez , (20)
Dernier
Dernier (20)
Jack Tuszynski Accelerating Chemotherapy Drug Discovery with Analytics and High Performance Computing
- 1. Jack Tuszynski Cross Cancer Institute Department of Physics University of Alberta Edmonton, Canada http://www.phys.ualberta.ca/~jtus “Accelerating Chemotherapy Drug Discovery with High Performance Computing and Analytics”
- 3. Modern Drug DevelopmentModern Drug Development Success Rate 1:100,000 !Success Rate 1:100,000 ! 00 22 44 66 88 1010 1212 1414 1616 DiscoveryDiscovery Preclinical testingPreclinical testing Phase IPhase I Phase IIPhase II Phase IIIPhase III ApprovalApproval Post marketPost market 100,000100,000 100100 55 11 Time in years Cost $1B
- 4. Identify disease Isolate protein Find drug Preclinical testing GENOMICS, PROTEOMICS & BIOPHARM. HIGH THROUGHPUT SCREENING MOLECULAR MODELING VIRTUAL SCREENING COMBINATORIAL CHEMISTRY IN VITRO & IN SILICO ADME MODELS Potentially producing many more targets and “personalized” targets Screening up to 100,000 compounds a day for activity against a target protein Using a computer to predict activity Rapidly producing vast numbers of compounds Computer graphics & models help improve activity Tissue and computer models begin to replace animal testing VIRTUAL SCREENING MOLECULAR MODELING The Evolution in Drug Design and Development
- 5. 5 Integration of biological dataIntegration of biological data impacts drug developmentimpacts drug development information stored in the genetic code (DNA)information stored in the genetic code (DNA) protein sequencesprotein sequences 3D structures of biomolecules3D structures of biomolecules experimental results from various sources (kd, IC50,experimental results from various sources (kd, IC50, expression)expression) clinical dataclinical data patient statisticspatient statistics scientific literaturescientific literature
- 6. 6 ……and leads toand leads to computational explosioncomputational explosion An avalanche of data:An avalanche of data: SequencesSequences Functional relationsFunctional relations StructuresStructures This requiresThis requires computationalcomputational approachesapproaches • 100’s of completed genomes • 1000’s of known reactions • 10,000’s of known 3D structures • 100,000’s of protein-ligand interactions • 1,000,000’s of known proteins & enzymes • Decades of biological/chemical know-how • Computational & Mathematical resources The Push to Systems Biology
- 7. 77 Key areas ofKey areas of bioinformaticsbioinformatics organisation of knowledge (sequences, structures, functional data) e.g. homology searches
- 8. Specifically for drug discovery: PDB : 50,000 proteins + homologs 1500 targets (human proteins) Approx. 400 (80 in cancer) utilized Orange Book: 1800 medicinal drugs Drug Bank: 4900 drugs Cancer chemotherapy drugs: 103 Protein-drug interactions but also Protein-protein interactions
- 9. Molecular Targets:Cancer Cell NetworkMolecular Targets:Cancer Cell Network A very complex but algorithmic system Based on a lock-and-key principle We will find keys to all these locks by 2061
- 10. CANCER CHEMOTHERAPY DRUGS Approximately 100 standard chemotherapeutic drugs: 1)Alkylating agents: Genotoxic (20-25) 2) Plant alkaloids: Inhibition of mitosis (10-15) 3) Antimetabolites: Inhibition of base synthesis (15-20) 4) Antibiotics: Derived from Streptomyces (10-15) 5) Targeted antibodies: Bind cell surface receptors (5-10) 6) Hormones: Inhibit or stimulate hormone signaling (15-20) 7) Directly targeting small molecules 8)Other indirect effects: Angiogenesis or immune modulators (10-15) Number of current chemotherapy targets: 101 Number of chemotherapy drugs: 102 Potential Targets (Pharmacogenomics): 103 Paclitaxel Cisplatin Methotrexate Trastuzumab Imatinib Tamoxifen Doxorubicin Bevacizumab
- 11. G2 M G1 S G0 tyrosine kinases DNA synthesis topoisomerase I CDK2 tubulin polymerisation/ depolymerisation Vinca alkaloids* taxol/taxotere halichondrin* spongistatin* rhizoxin* cryptophycin sarcodictyin eleutherobin epothilones discodermolide D-24851 ? dolastatin* combretastatin* camptothecin CDK4 flavopiridol (R)-roscovitine (CYC202) paullones, indirubins gleevec iressa OSI774 hydroxyurea cytarabine antifolates 5-fluorouracil 6-mercaptopurine nitrogen mustards nitrosoureas mitomycin C CDK1 Chk1 Chk2 UCN-01, SB-218078 debromohymenialdisine isogranulatimide AhR actin kinesin Eg5 monastrol ecteinascidin 743 podophyllotoxin,doxorubicin etoposide, mitoxantrone topoisomerase II ATM/ATR R115777 SCH66336 ROCK Y-27632 CDC25 DF203 FK317 HMGA Plk1 Aurora wortmanni n caffeine ODC/SAMDC Pin1 GSK-3 Cdc7 nucleotide excision repair Raf cytochalasins latrunculin A scytophycins dolastatin 11 jasplakinolide paullones, indirubins (R)-roscovitine (CYC202) paullones, indirubins BAY-43-9006 fumagillin,TNP-470 PRIMA-1, pifithrin a rapamycin mTOR/FRAP PS-341 proteasome bryostatin, PKC412 PKC histone deacetylasetrichostatin, FK228 HSP90geldanamycin, 17- AAGATK, MAFP cytosolic phospholipase A2 hexadecylphosphocholin e phospholipase D CT-2584 choline kinase MEK1/Erk-1/2 PD98059, U0126 menadione (K3) farnesyl transferase phosphatasesokadaic acid, fostreicin, calyculin A Wee1 PD0166285 polyamine analogues Pin1 p53/MDM2 Source: Cell cycle laboratory, L. Meijer, Roscoff, France ~80 drugs and drug candidates Cancer chemotherapy is based on cell cycle arrest
- 12. CAUSES OF FAILURE IN DRUG DEVELOPMENT ADME ANIMAL TOXICITY LACK OF EFFICACY ADVERSE EFFECTS IN HUMANS More than 50% of this failure can be predicted computationally in 2011 In 2061: six sigma will be achieved in silico
- 13. WET LAB: High-throughput screening (HTS)WET LAB: High-throughput screening (HTS) Experimental techniqueExperimental technique 384-well microplates, florescence-based detection &384-well microplates, florescence-based detection & desktop robotsdesktop robots Up to 1M compounds per targetUp to 1M compounds per target DRY LAB: Virtual screening (VS)DRY LAB: Virtual screening (VS) Ligand-based methodsLigand-based methods 2D structures, substructures, fingerprints2D structures, substructures, fingerprints Volume/surface matchingVolume/surface matching 3D pharmacophores, fingerprints3D pharmacophores, fingerprints Receptor-based methodsReceptor-based methods DockingDocking Even 100B compounds per target triedEven 100B compounds per target tried Receptor flexibility
- 14. OUR 1024-PROCESSOR HPC CLUSTER WE ALSO USE 500 PROCESSORS FROM WEST-GRID AND SHARCNET
- 15. Target-Protein Structure MRECISIHVGQAGVQIGNACWELYCLEHGIQPDGQMPSDKTIGGGDDSFNTFFSETGAGKHVPRAVFVDLEPTV IDEVRTGTYRQLFHPEQLITGKEDAANNYARGHYTIGKEIIDLVLDRIRKLADQCTGLQGFSVFHSFGGGTGSGFT SLLMERLSVDYGKKSKLEFSIYPAPQVSTAVVEPYNSILTTHTTLEHSDCAFMVDNEAIYDICRRNLDIERPTYTNL NRLIGQIVSSITASLRFDGALNVDLTEFQTNLVPYPRGHFPLATYAPVISAEKAYHEQLSVAEITNACFEPANQMV KCDPRHGKYMACCLLYRGDVVPKDVNAAIATIKTKRTIQFVDWCPTGFKVGINYEPPTVVPGGDLAKVQRAVCM LSNTTAIAEAWARLDHKFDLMYAKRAFVHWYVGEGMEEGEFSEAREDMAALEKDYEEVGVDSVEGEGEEEGE EY Primary: amino acid sequence Secondary: -helix and -α β sheet Tertiary: 3D-folding Quaternary: multimeric arrangement
- 16. Molecular Dynamics • Treats molecules classically: – Point charges and masses – Spring-like bonds – Numerical integration of equations of motion
- 17. Drug binding sites in tubulin Of the more thanOf the more than 100100 approvedapproved cancer chemotherapy drugs oncancer chemotherapy drugs on the market, approximately 15%the market, approximately 15% target tubulin directly.target tubulin directly. None are specific for cancerNone are specific for cancer cells, hence associated sidecells, hence associated side effectseffects
- 18. Drug / Ligand Protein Drug ActionDrug Action: Inhibition of Protein-: Inhibition of Protein- Protein InteractionsProtein Interactions Cavity Cavity Cavity
- 19. The computational toolboxThe computational toolbox The three-fold way:The three-fold way: rational design andrational design and in silicoin silico testing of derivatives of knowntesting of derivatives of known agentsagents brute-force computational search using existing librariesbrute-force computational search using existing libraries (pharma-matrix)(pharma-matrix) De novo design from common pharmacophores for bestDe novo design from common pharmacophores for best space filling propertiesspace filling properties a pocketome data banka pocketome data bank Reverse docking allows to predict side effectsReverse docking allows to predict side effects
- 20. How Do We Solve Our Puzzles?
- 21. ContentsContents Compound dataCompound data sourcessources (PubChem, Zinc, NCI, SciFinder(PubChem, Zinc, NCI, SciFinder ~65M compounds)~65M compounds) Drug dataDrug data sourcessources (DrugBank, Orange Book, CMC, WDI,(DrugBank, Orange Book, CMC, WDI, MDDR ~ 250 k drugs)MDDR ~ 250 k drugs) Molecular dataMolecular data toolkitstoolkits (OpenEye, Open Babel)(OpenEye, Open Babel) Computational MethodsComputational Methods (MM, MD, QMMM)(MM, MD, QMMM) Molecule file formatsMolecule file formats (PDB, Smilies )(PDB, Smilies ) DockingDocking (Autodock, Dock)(Autodock, Dock) ParallelParallel (Dovis)(Dovis)
- 22. Pharma-matrix apps:Pharma-matrix apps: eRxeRx 100 million targets (100,000 proteins x 100 pockets x 10 mutants):100 million targets (100,000 proteins x 100 pockets x 10 mutants): pocketomepocketome 100 billion chemical compounds100 billion chemical compounds 10101919 potential interactions (filtering)potential interactions (filtering) Hand-in-glove match by brute computational screeningHand-in-glove match by brute computational screening pharmagooglepharmagoogle
- 23. Pocketome generation (pocket clustering) 104 clusters 104 pockets in a cluster Docking (1012 calculations within blocks) Docking (1012 calculations within blocks)
- 25. Personalized eDx and eRx in a few decades a personal genome will cost $10 and will be our ID at birth included in our eRx app
- 26. The Virtual Human:The Virtual Human: Multi-Scale ModelingMulti-Scale Modeling lobule liver whole body hepatocyte Drug molecules Interaction matrix
Notes de l'éditeur
- Let’s take a small part of the picture here. There many areas of research in this picture. In this presentation I will focus on these two subjects which are basically our research group interest.
- The problem is even worse. Proteins may undergo post-translational modifications, associations with other molecules or prosthetic groups, and formation of multimeric complexes. In medicine, a prosthesis (plural prostheses) is an artificial extension that replaces a missing body part. It is part of the field of biomechatronics, the science of fusing mechanical devices with human muscle, skeleton, and nervous systems to assist or enhance motor control lost by trauma, disease, or defect.
- Cisplating – crosslinks DNA Paclitaxel – binds tubulin and inhibits chromasome seggregation Methotrexate – inhibits folate synthesis Doxorubicin – DNA intercelator Trastuzumab – HER2 Tamoxifen – estrogen antagonist Imatinib Gleevec – tyrosine kinase inhibitor Bevacizumab - binds VEGF and inhibits angiogenesis
- -h-bonds -salt-bridges -hydrophobic, hydrophilic
- Here is a close up view of the secondary strucutre of tubulin as it is observed in a protofilament. Several drug binding sites have been characterized crystallographically. Shown as VDW spheres are three of the best characterized tubulin binding drugs, Taxol, vinblastine and colchicine. There are currently about 100 approved chemotharapy drugs on the market, several of which are monoclonal antibodies. Of these drugs, about 15% specifically target tubulin in some way reducing or increasing their dynamicity. Unfortunately, none of the tubulin binding chemotherapy drugs are specific for cancer cells. I will folcus on colchcine for this talk
- When the structure of the target protein is known, the drug discovery process usually follows a well-established procedure shown schematically in Figure 1. Virtual screening techniques are applied early during the docking protocol to reduce the size of large compound libraries. Initially, libraries are ‘‘pre-filtered’’ using a series of simple physicochemical descriptors to eliminate compounds not expected to be suitable drugs. Pharmacophore analysis, neural nets, similarity analysis, scaffold analysis, Lipinski’s rule of five. This procedure, which reduces the size of the library to a group of molecules more likely to bind the target receptor, is known as enrichment. Once an optimum library has been produced, molecules are docked to the target receptor to reduce further the number of candidates. This initial screening makes use of fast, but not very accurate, ranking functions to evaluate the relative stability of the docked complexes. The selected candidates, usually a few hundred, are subject to further docking experiments using more sophisticated scoring functions
- When the structure of the target protein is known, the drug discovery process usually follows a well-established procedure shown schematically in Figure 1. Virtual screening techniques are applied early during the docking protocol to reduce the size of large compound libraries. Initially, libraries are ‘‘pre-filtered’’ using a series of simple physicochemical descriptors to eliminate compounds not expected to be suitable drugs. Pharmacophore analysis, neural nets, similarity analysis, scaffold analysis, Lipinski’s rule of five. This procedure, which reduces the size of the library to a group of molecules more likely to bind the target receptor, is known as enrichment. Once an optimum library has been produced, molecules are docked to the target receptor to reduce further the number of candidates. This initial screening makes use of fast, but not very accurate, ranking functions to evaluate the relative stability of the docked complexes. The selected candidates, usually a few hundred, are subject to further docking experiments using more sophisticated scoring functions