JAMA+ AI Conversations AI for Drug Discovery
Nov 6, 2025
Ajamete Kaykas, Chief eXploration Officer at Insitro, shares his expertise on harnessing AI for drug discovery. He discusses the innovative ClinML, CellML, and ChemML platforms that combine human data and cellular perturbations. Ajamete highlights the importance of integrating public datasets like UK Biobank with proprietary data. He addresses the gradual impact of AI in the field and emphasizes its role as an assistant to scientists. Finally, he advocates for cross-disciplinary training to foster the next generation of biologist-data scientists.
AI Snips
Chapters
Transcript
Episode notes
Integrated Multi‑Platform Discovery Engine
- Insitro runs three integrated platforms: ClinML for human data, CellML for large-scale cell perturbations, and ChemML for DNA-encoded chemistry.
- Combining these platforms creates a discovery engine linking human genetics to cellular models to small-molecule discovery.
Integration Beats Ad‑Hoc AI Patching
- Insitro prioritized building an AI/ML-first company with integrated data pipelines and a common database from day one.
- Tight integration of bilingual biologists and computational scientists accelerates translation and avoids silo problems.
Create Controlled Data, Capture Metadata
- Generate your own controlled, high-quality experimental data when public data lack detail or rigor.
- Capture complete metadata and rigorous protocols so ML models can be properly interpreted and trusted.