Download to read offline
Uploaded byChris Southan
PubChem as a source of systems biology perturbagens
The document discusses the utility of chemical perturbation in systems biology, focusing on the advantages of using various drugs and probes available in the PubChem database. It highlights the growing chemical toolbox for researchers and the importance of reproducibility and mechanistic insights in perturbation studies. The author emphasizes the need for better indexing of probes and the benefits of integrating multiple data sources for effective target engagement and drug discovery.
More Related Content
![[4] drug discovery-lect-1](https://cdn.slidesharecdn.com/ss_thumbnails/4-drugdiscovery-lect-1-130821015918-phpapp02-thumbnail.jpg?width=640&height=640&fit=bounds)
Similar to PubChem as a source of systems biology perturbagens
More from Chris Southan
PubChem as a source of systems biology perturbagens
- 1. The utility offor selecting drugs and probes as systems biology perturbagens Christopher Southan Modelling Biological Systems Symposium, Imperial College London, 26 March 20191 https://sites.google.com/view/tw2informatics/home
- 2. Outline • Chemical perturbation •PubChem content and coverage • Targets • Drugs • Probes • Looking at BAY 7598 • Getting into PubChem • Conclusions 2
- 3. Advantages of chemicalpeuterbation of biological systems • Dose response, rapid onset, and time course measurement • Reversable by washout > retest by pulsing • Should be robustly reproducible • Broad choice of controls (e.g. different binding sites, potencies, kinetics, chemotypes, SAR including inactive analogues) • In vitro and In cellulo target engagment/mechanistic validation • Protein/Protein interaction inhibitors • Option of multiple pathway/network intervention testing • Omic profiling of effects (e.g. transcripts, imaging, metabolites, proteomics) • Orthogonal hypothesis testing viaCRISPR, RNAi, KO, antibody ect • Peptides, nucleotides, antibodies or endogenous metabolites can be tried • Druggable genome target coverage increasing (e.g. PROTACs) • Can use compounds with unkown mechanism (e.g. phenotypic screening hits) • Not restricted to proteins (e.g. RNA-small molecule interactions) • ~ 3 million bioactive compounds available in PubChem • Analogue expansion by synthetic chemistry 3
- 4. TheTriumvirate: Substance, Compound, BioAssay 4 •SIDs can be biololecules (e.g. large peptides and antobodies, no images • CIDs merge SMILES strings < 1000 atoms, to unique InChIs • Average SID:CID ~ 2.6, drugs ~50, aspirin (CID 244) 307 plus 1563 mixture SIDs • CIDs 4.7 % mixtures
- 5. PubChem CID growth2005 - 2018 5
- 6. CID statistics overview (March 2019) 6
- 7. Potential perturbagen coveragein PubChem • 80 million rule-of-five chemical space • 3.4 million tested > 1.25 million active in BioAssay • 1.0 million assays > 240 million bioactivities • Activities extracted from ~ 100K papers • 12,000 protein targets • ~ 8000 PDB ligands • 12,000 with BioSytems pathway mappings • 9500 drugs, ~ 1500 FDA approved • BindingDB 20,500 cpds, 260 patents, 42,000 activities (just 2019) • Activity data not in PubChem but in patents ~2.5 million • Probes, full leads and develpment cpds unkown (not cleanly tagged) 7
- 8. Curated sources ofhuman targets 8 • All four are PubChem submitting sources for linked chemistry • Most extracted from publications • They all have UniProt cross-references • Numbers shown are for Human Swiss-Prot • Represent a maximal and consensus druggable genome
- 9. Drugs vs leadsvs probes vs tools: Drugs • Approved drugs have accumulated data on systems perturbations and molecular mechanisms of action • However, the primary human target number is below 400 and little may be known on secondary targets and (systems) polypharmacology • Optimisation of ADMET drug parameters may reduce desirable perturbagen properties such as potency, Kd in vitro and in cellulo • However the same ADMET optimisation means they can go straight into cells, organoids, rodents, zebrafish or other model systems • Chemical vendor and patent SAR coverage are usefully high for drugs • Drugs are designed for patient efficacy but not to provide systems biology mechanistic insights per se • Past and present development compounds (~ 15,000 +) cover a much wider range of targets but are difficult to select cleanly in PubChem 9
- 10. Selecting approved drugsfrom Guide to Pharmacology 10 9526 SIDs > 1509 approved drugs > mapping to 1321 CIDs
- 11.
- 12.
- 13. Probes, repurposing andtools • Probes are designed to modulate a target for mechanistic and system insights rather than therapeutic utility in first instance • Probe guidelines published on potency, specificity and activity in cells • Interesting differences between the NIH first generation probes circa 1990- 2010 and the later generation from SGC and others in last few years • Repurposing compounds from AstraZeneca, Boehringer and other companies also potentially useful as perturbagens • Tool compound is a broader term for anything with useful specificity that might not be strictly a drug or a probe (e.g. spider venoms against ion channels) 13
- 14. Useful external probesource for ”mapping in” to PubChem 14
- 15. Tracking a probeexample into PubChem 15
- 16. Not so easyto get the structure 16
- 17. The bad news: PubChem has it from a Bayer patent but no BioAssay data and not named as BAY 7598 17
- 18. The good news: MMP12 is indexed in PubChem with 7 Guide to Pharmacology curated inhibitors 18
- 19. You too canget into PubChem 19 If your perturbation experiments are reproducible then put the results into PubChem BioAssay (to be FAIR) https://cdsouthan.blogspot.com/2017/08/getting-into-pubchem-again.html
- 20. Conclusions • As thede facto global nexus for bioactive chemistry PubChem has become an essential resource for perturbagen choices and cross-checks • Between drugs, development compounds, leads, probes and patent SAR sets there is a increasingly wide chemical toolbox choice • PubChem synergises with the stand-alone sources integrated into it such as ChEMBL, Guide to Pharmacology, DrugBank, BindingDB and DrugCentral • Note these may offer easier entry points but follow structures back into PubChem for source checking, BioAssay assessment, neighbour assessment and patent links • UniProt offers an alternative entry point from the chemically modulatable protein target side and their cross-reference intersecting is very powerful • Despite their increasing importance collective indexing of probes is poor • Like all such sources, PubChem it has navigation quirks, caveats, submitter quality issues, gotchas and cheminformatic challenges 20
- 21.
- 22. Further PubChem tipsand tricks 22 https://www.slideshare.net/cdsouthan/presentations https://cdsouthan.blogspot.com/