Drug candidates fail in the clinic not because they miss their target, but because of ADMET — absorption, distribution, metabolism, excretion, and toxicity. Pat Walters has spent 30 years watching this happen. Now, as Chief Scientist at OpenADMET, he’s building the open data infrastructure to fix it.
Pat argues that data — not better algorithms — is the real bottleneck in applying AI to drug discovery, and explains why you can’t just pull reliable ADMET data from the literature. He walks through OpenADMET’s approach: generating large, consistent, publicly available datasets; running blind prediction challenges (370 groups participated in their first); and integrating structural biology to move from black-box models to mechanistic understanding.
Links
- OpenADMET: https://openadmet.org/
- Practical Cheminformatics: https://patwalters.github.io