For each target, a per-target QSAR model (or, for targets with few known ligands, a 2D interaction pharmacophore) ranks an 8.4M in-stock library; a diversity-selected ~10k shortlist is docked with Uni-Dock into one curated structure and rescored by gnina CNNaffinity (or Vina for nuclear receptors). The top 1000 per target are kept.
Post-filtering: physically clashing poses (Vina > 0) and PAINS are removed; 12 targets whose pockets are too small for drug-like matter are excluded; zinc-enzyme and aminergic shortlists are restricted to the appropriate zinc-binding-group / basic-amine chemotype. Scores are relative ranking scores, not calibrated affinities.
Each target's top 1000 is matched by scaffold (InChIKey skeleton) against its ChEMBL actives (pChEMBL ≥ 7). A target “recovers” a known active when that active's scaffold appears in its top 1000. Because most known actives are research compounds not purchasable in the library, this is a conservative signal: recovery means the docking floated a genuine active to the top of the in-stock matter that was available.
| Target | Class | # known | # recovered | best rank | top-100 sim |
|---|
“top-100 sim” = median Morgan(2,2048) Tanimoto of the top-100 hits to their nearest known active (how active-like the ranked matter is). Sort any column by clicking its header.