OpenADMET / Avoid-ome
Open-science initiative creating pre-competitive mechanistic ADMET datasets targeting the 'Avoid-ome' — proteins acting as anti-targets. Uses high-throughput structural biology and active learning with community challenges.
Composite
53.6
Experimental validation
Retrospective
Stages
Lead ID / ADMETIND-enabling
Modalities
small molecule
Task types
classificationregression
Size
proteins: 50
molecules: 100,000
splits: {'train': 0, 'val': 0, 'test': 0}
molecules: 100,000
splits: {'train': 0, 'val': 0, 'test': 0}
License
CC-BY
First release
2026-05-25
Last updated
2026-05-25
Official site
Leaderboard
→ leaderboard
Dataset
→ dataset
Code / GitHub
→ repository
HuggingFace
→ HF
Paper
Mapping the avoid-ome: a systematic open-science approach to predictive ADMET · James S. Fraser, Steven Edgar, L. Naomi Handly, Sriram Kosuri, John D. Chodera, Mark Murcko, W. Patrick Walters · 2026 · paper · doi:10.1038/s41467-026-73410-8 · 0 citations
Flags
none
Experts
—
Groups
—
Hosted by
—
Related benchmarks
Rubric (7-criterion)
rigor
5
coverage
3
maintenance
4
adoption
2
quality
4
accessibility
3
industry_relevance
5
Notes
Nat Comms perspective (May 2026) by top structural biology/comp chem leaders (Fraser, Chodera, Murcko, Walters). Proposes mechanistic ADMET datasets grounded in structural 'ground truth'. Still early — datasets forthcoming — but the consortium authority and Nature publication warrant inclusion. High industry relevance given Relay/MSKCC/UCSF backing.