ML & AI in Drug development: the hidden part of the iceberg
0 likes215 views
This document summarizes the challenges of applying machine learning and artificial intelligence to drug development. It discusses how drug development is a long and complex process involving identifying disease targets, developing drug candidates, and testing through clinical trials. It then explains that biology is complex, data is often incomplete or biased, and there is a lack of labeled examples, making application of AI difficult. However, areas that could benefit include using AI to subtype diseases, interpret medical images like tumors, and build knowledge graphs to discover new insights. More and better quality data, along with focus on interpretability and engineering practices, are needed to further progress in this area.
ML & AI in Drug development: the hidden part of the iceberg
- 1. ML & AI in Drug development: the hidden part of the iceberg Paul Agapow Oncology R&D July 2021 Public
- 2. Disclosure • No conflicts of interest • Does not reflect official AZ thought or projects • Based on experience in current & previous positions • ML&AI / Health Informatics @AZ • Data Science Institute @ICL • Bioinformatics @Health Protection Agency (UK) … 2
- 3. What is drug development, how does it work? Agenda 3 Why ML & AI is difficult in pharma Where ML & AI can be powerful in pharma and what we need to do 1 2 3
- 4. How we make drugs 1
- 5. Clinical trials Identifying and understanding disease, unravelling the molecular machinery, pinpointing targets Drug development is a long & complex process 5 Pathophysiology Developing molecules that can be synthesized and delivered safely to the target Drug candidates Testing & optimizing trials, dissecting failures and successes, tracking adverse events, seeking regulatory approval Who gets the drug, how is it re-imbursed, tracking long-term adverse events Post-approval
- 6. 6 • ~ $2B and 10 years to develop & launch a drug • The “valley of death”: most candidate drugs will fail The tough maths of drug development ePharmacology.hubpages.com
- 7. Why ML & AI is difficult in pharma 2
- 8. 10 June 2021 8 “AI will not replace drug hunters, but drug hunters who don’t use AI will be replaced by those who do.” -Andrew Hopkins, CEO Exscientia
- 9. 9
- 10. 12 July 2021 10 The complexity of biomedicine: About 50 trillion cells of 200 types Each cell has 23 pairs of chromosomes In total 6.4 billion basepairs (positions) Organised into about 18,000 genes (Or maybe more like 40,000 genes) Genetic material elsewhere in the cell Epigenetic modification 1 million different types of molecules Lifestyle & history Exposure & environment Immune system repertoire & priming … Of which we know only a fraction
- 11. Why? • Biology is outrageously complex • Data is frequently biased, irregular, incomplete, in different formats • Biomedicine is a label desert • As a consequence: • Advances are throttled by domain knowledge • How to represent & analyse complex domain • Suitable data is often scarce 11
- 12. The classic analytical tension 12 What we need to solve What we tend to solve Easy things Available, ideal data Ground truth Simplify “Interesting” “Table-land” Useful things Incomplete messy data Unclear biological reality Uncertain findings Needful “Network-land”
- 13. Where can ML & AI be powerful in pharma? 3
- 14. 14 Radiology & imaging widely used in healthcare • Capture important & difficult to abstract data –E.g. presence, size, shape of tumor • Radiologists –Never enough of them –Rushed –Frequently wrong • But AI is good at interpreting images …
- 15. 15 Not just X-rays & MRI but microscopes Cancers are associated with certain proteins, which traditionally have to be stained & examined visually. Deep learning can automatically do this for us Slide stained for PD-L1 expression Cells that were automatically detected using AI
- 16. 16 Precision medicine: subtypes of diseases & patients Type 2 Diabetes Topology based Patient-Patient network, identify distinct subtypes of T2D Dudley et al. Sci. transl. Med, 2015 COPD • Transform patients into sequences of diagnosis codes • Look for over-represented temporal pairs of codes • Collapse pairs into trajectories of diagnoses • Combine similar trajectories with graph similarity Brunak et al. Nature Coms. 2016
- 17. 10 June 2021 17 • A lot of biomedical knowledge is associative or relational & multimodal • Knowledge graphs / GCNs help us to capture and analysis • Have been used to propose new drugs and patient subtypes
- 18. Software as a Medical Device 18 • For: • Diagnosis • Monitoring • Interpretation • Prognosis • Priorotorization … • Tough regulatory environment
- 19. 19 How to manage and interoperate? Target ID Target Validation Discovery Pre-Clinical Clinical Commercial Post Marketing Surveillance Genetic & Genomic Data Patient-Centric Data Sensors & Smart Devices Interactive Media Healthcare Information network Market Data We need more data
- 20. Interpretability (etc.) is vital • But what actually is interpretability? • May feedback to inspire mechanistic research, but … • Essential for: • a smoke test, validation • check for bias • communication • Likewise calibration • Important to understand how (un)sure we are 20
- 21. Good (engineering) practices & production quality is vital 21
- 22. Takeaways Drug development is a enormously complex process Although attractive, ML & AI are often hindered by the nature of the data Areas of definite value include subtyping, imaging & knowledge graphs Gathering & more and wider data and better engineering is key to further progress 22
- 23. Click to enter title here Some light reading 23 Academic Press (2021) Editor: Ashenden
- 24. Looking for work? • If you are driven by science and passioned about improving lives, why not look at a job in pharma? • Example jobs at AstraZeneca - visit our careers website for much, much more: • Principal Data Scientist • Associate Director Imaging & AI - Imaging & Data Analytics • Knowledge Graph Engineer • Data Sciences & AI Graduate Programme 24
- 25. Confidentiality Notice This file is private and may contain confidential and proprietary information. If you have received this file in error, please notify us and remove it from your system and note that you must not copy, distribute or take any action in reliance on it. Any unauthorized use or disclosure of the contents of this file is not permitted and may be unlawful. AstraZeneca PLC, 1 Francis Crick Avenue, Cambridge Biomedical Campus, Cambridge, CB2 0AA, UK, T: +44(0)203 749 5000, www.astrazeneca.com 25