PoseBusters

Physics-aware eval of docking/co-folding poses — 19 checks catching chemically impossible outputs.

Composite

97.0

Experimental validation

Retrospective

Stages

Hit ID

Modalities

small-moleculeprotein-general

Task types

pose-validation

Size

complexes: 428
checks_per_pose: 19

License

BSD-3-Clause

First release

2023-08

Last updated

2025-02

Official site

Leaderboard

Dataset

Code / GitHub

HuggingFace

→ HF

Paper

PoseBusters: AI-based docking methods fail to generate physically valid poses or generalise to novel sequences · Buttenschoen M, Morris GM, Deane CM · 2024 · paper · doi:10.1039/D3SC04185A · 360 citations

Flags

none

Experts

Charlotte Deane, Martin Buttenschoen

Groups

Oxford OPIG (Deane Lab)

Hosted by

PoseBusters Evaluation Suite

Related benchmarks

PLINDER, PINDER, CASF-2016

Rubric (7-criterion)

rigor

coverage

maintenance

adoption

quality

accessibility

industry_relevance

Notes

Exposed major failure modes in AlphaFold-Multimer/DiffDock/RFAA. Default pharma filter.

← Back to all benchmarks

Compare:

Open comparison →