Value chain — stage 01
Discovery & preclinical
Target identification, generative chemistry, structure prediction, ADMET and the new approach methodologies replacing animal studies — what the models can and cannot do yet.
- Target identification & validation
- Generative & de-novo design
- Structural biology
- ADMET & translational safety
- Lab automation
In depth
2 articles
Specific problems, specific answers, each written against a named source and dated.
Phase 1 Improved. Phase 2 Did Not.
Generative chemistry optimises molecular properties, which Phase 1 tests; target validity, which Phase 2 tests, remains almost completely untouched today.
Discovery · 11 min readWhy Predictive Models Collapse on Your Own Chemistry
Temporal splits, applicability domain and genetic evidence: the unglamorous baselines that absorption, distribution and target identification models must actually beat.
Discovery · 13 min read