ICIC Berlin October 2012 - Dr. Haxel

The  ChEMBL  Database    

ICIC  2012  Berlin,  Germany  October  2012  

 John  P.  Overington  

EMBL-­‐EBI    

[email protected]  

Drug-like compounds

Chemical  Space  All compounds Available compounds

Only certain molecules have features consistent with good pharmacological properties

Druggable targets

Target  Space  

Only certain targets have binding sites capable of ligand efficient binding of drug-like ligands

All targets Available targets

Accessible  Pharmacological  Space  

Available compounds for

target but non-drug-like

Drug-like compounds but no complementary

targets

Druggable targets but no

complementary compounds

Druggable targets and complementary

compounds

Pre

sent

ed to

P&

G, C

inci

nnat

i, A

pril

2005

, © 2

005

Inph

arm

atic

a Lt

d.

All reasonable molecules 1020

All reasonable proteins

106

Screened proteins 103

Screened molecules 107-8

ChEMBL

Chemogenomics  

Exploration of bioactivity space at genomic scale Structure Activity Relationship (SAR)

Drugs 103

Drug targets 102 Drugs

ChEMBL  Database  

•  hKp://www.ebi.ac.uk/chembl  •  Funded  by  a  Strategic  Award  from  the  Wellcome  Trust  •  World’s  largest  primary  source  of  Open  pharmacology/drug  

discovery  data  –  Contains  syntheTc  small  molecules,  natural  products  and  biologicals  –  Strong  integraTon  and  annotaTon  of  chemical  and  biological  data  –  OSINT  approach  to  data  gathering  –  Tight  integraTon  with  other  EBI  resources  

•  Ensembl,  1000  Genomes,  UniProt,  PDBe,  ArrayExpress,  Atlas….  –  Data  sharing  agreements  in  place  with  key  public  resources,  e.g.  PubChem  

•  Open  Data  –  CC-­‐BY-­‐SA  licence  •  Free  downloads,  secure  private  searching,…  •  REST  web  service  API  

Target  Discovery  

Lead  Discovery   Lead  OpTmizaTon  

Preclinical  Development  

Phase  1   Phase  2   Phase  3   Launch  (Phase  4)  

Drug  Discovery  

~1,400,000  compound  records  >10,000,000  bioacTviTes  ~46,000  abstracted  papers  ~9,000  targets  

~12,000  clinical  candidates  

~1,600  drugs  

•   Target  idenTficaTon  •   Microarray  profiling  •   Target  validaTon  •   Assay  development  •   Biochemistry  •   Clinical/Animal  disease  models  

•   High-­‐throughput  Screening  (HTS)  •   Fragment-­‐based  screening  •   Focused  libraries  • Screening  collecTon  

•   Medicinal  Chemistry  •   Structure-­‐based  drug  design  •   SelecTvity  screens  •   ADMET  screens  •   Cellular/Animal  disease  models  •   PharmacokineTcs  

•   Toxicology  •   In  vivo  safety  pharmacology  •   FormulaTon  •   Dose  predicTon  

PK  tolerability   Efficacy  

Safety  &  Efficacy  

IndicaTon  discovery,  repurptg  &  expansion  

Med.  Chem.  SAR   Clinical  Candidates   Drugs  

Discovery   Development   Use  

ChEMBL content

Only  ~1%  of  Genome  is  a  Drug  Target  

Drug  Approvals  

FDA  Approved  Drugs  

NFκB  Pathway  –  key  control  mechanism  for  inflammaTon  

Affinity  of  Drugs  for  their‘Targets’  

Ki,  Kd,  IC50,  EC50,  &  pA2  endpoints  for  drugs  against  their‘efficacy  targets’  

2   3   4   5   6   7   8   9   10   11   12  0  

50  

100  

150  

200  

250  

300  

350  

400  

Freq

uency  

-­‐log10  affinity  

10mM   1mM   100mM   10mM   1mM   100nM   10nM   1nM   100pM   10pM   1pM  

Overington,  et  al,  Nature  Rev.  Drug  Discov.  5  pp.  993-­‐996  (2006)  Gleeson  et  al,  Nature  Rev.  Drug  Discov.  10  pp.  197-­‐208  (2011)  

Clinical  Candidates  

•  CollecTon  of  clinical  development  candidates  –   Contains  ~12,000  2-­‐D  structures/sequences  

•   EsTmated  size  ~35-­‐45,000  compounds  

–   Work  in  progress  •  e.g.  Protein  kinases,  393  disTnct  clinical  candidates  

Different  Types  of  Drugs  

Pharma  Industry  ProducTvity  File  RegistraTon  number  vs.  USAN  date  

0  

100,000  200,000  300,000  400,000  500,000  600,000  700,000  800,000  

1960   1965   1970   1975   1980   1985   1990   1995   2000   2005   2010  

Phase 2b date

~Discovery date

Overington,  unpublished  

Pharma  Industry  ProducTvity  

0  

10  

20  

30  

40  

50  

60  

70  

1-­‐  100,000  

100,001-­‐  200,000  

200,001-­‐  300,000  

300,001-­‐  400,000  

400,001-­‐  500,000  

500,001-­‐  600,000  

600,001-­‐  700,000  

700,001,  800,000  

File  registraTon  number  range  

64  USANs/100,000  compounds  

1.9  USANs/100,000  compounds  

16  Drugs/100,000  compounds  

0.4  Drugs/100,000  compounds  

Large  Pharma  needs  on  average  to  synthesize  and  test  ~250,000  compounds  for  each  launched  drug  

Overington,  unpublished  

Patent  and  PublicaTon  Lag  

IBM  Patent  data  and  ChEMBL  

Clinical  Candidates  

What  Is  the  ChEMBL  Data?  

SAR  Data  

Compound  

Assay  

Ki=4.5  nM  

>Thrombin   MAHVRGLQLPGCLALAALCSLVHSQHVFLAPQQARSLLQRVRRANTFLEEVRKGNLERECVEETCSYEEAFEALESSTATDVFWAKYTACETARTPRDKLAACLEGNCAEGLGTNYRGHVNITRSGIECQLWRSRYPHKPEINSTTHPGADLQENFCRNPDSSTTGPWCYTTDPTVRRQECSIPVCGQDQVTVAMTPRSEGSSVNLSPPLEQCVPDRGQQYQGRLAVTTHGLPCLAWASAQAKALSKHQDFNSAVQLVENFCRNPDGDEEGVWCYVAGKPGDFGYCDLNYCEEAVEEETGDGLDEDSDRAIEGRTATSEYQTFFNPRTFGSGEADCGLRPLFEKKSLEDKTERELLESYIDGRIVEGSDAEIGMSPWQVMLFRKSPQELLCGASLISDRWVLTAAHCLLYPPWDKNFTENDLLVRIGKHSRTRYERNIEKISMLEKIYIHPRYNWRENLDRDIALMKLKKPVAFSDYIHPVCLPDRETAASLLQAGYKGRVTGWGNLKETWTANVGKGQPSVLQVVNLPIVERPVCKDSTRIRITDNMFCAGYKPDEGKRGDACEGDSGGPFVMKSPFNNRWYQMGIVSWGEGCDRDGKYGFYTHVFRLKKWIQKVIDQFGE

 ED2=230  nM  

What  Is  the  ChEMBL  Data?  

Inhibition of human Thrombin

PTT (partial thromboplastin time)

ChEMBL  Target  Types  

Protein  complex  

e.g.  NicoTnic  acetylcholine  receptor   e.g.  Muscarinic  receptors   e.g.  DNA  

e.g.  Mitochondria  e.g.  Trachea  e.g.  HEK293  cells   e.g.  Drosophila  

e.g.  PDE5  

Protein   Nucleic  Acid                Protein  family  

Cell  line   Tissue   Sub-­‐cellular  frac>on   Organism  

Compound  Searching  

21

Spreadsheet  Views  

22

Ligand  Efficiency  

23

•  Ligand  efficiency  is  an  objecTve  measure  of  how  much  binding  energy  comes  from  each  atom  in  a  parTcular  interacTon  – Drugs  have  high  ligand  efficiency  

–  Every  atom  counts  – Need  to  avoid  affinity  from  lipophilicity  

Target  Class  Data  

Assay  Organism  Data  

Allosteric  Regulators  •  Allosteric  drugs  can  have  some  advantages  over  orthosteric  drugs  – SelecTvity  – Orthosteric  site  may  be  undruggable  

Allosteric/Orthosteric  sites  for  GPCRs  

hKp://www.chemblog.org