OpenBind EV-A71 Structure-Affinity Dataset

First open dataset from the OpenBind consortium (Oxford/Diamond Light Source). Dense structure-affinity dataset for Enterovirus A71 2A protease: 925 crystallographic binding events from 699 compounds with affinity measurements for 601 compounds. Designed for benchmarking structure-based AI docking, cofolding, and affinity prediction.

Composite
84.8
Experimental validation
Wet-lab confirmed
Stages
Hit IDLead ID / ADMET
Modalities
protein_structuresmall-molecule
Task types
binding_affinitypose_predictionvirtual_screening
Size
crystal_structures: 925
affinity_measurements: 601
License
CC-BY-4.0
First release
2026-05
Last updated
2026-05
Official site
→ project page
Leaderboard
→ leaderboard
Dataset
→ dataset
Code / GitHub
→ repository
HuggingFace
→ HF
Paper
OpenBind's first release: a structure-affinity dataset for structure-based AI · · 2026 · paper · 0 citations
Flags
experimental_validationcrystallography
Experts
Groups
Hosted by
Related benchmarks
CASF-2016, PDBbind, PoseBusters

Rubric (7-criterion)

rigor
5
coverage
2
maintenance
5
adoption
3
quality
5
accessibility
5
industry_relevance
5

Notes

One of the largest public single-target structure-affinity datasets. High-throughput crystallography at Diamond Light Source. Plans for more targets and blind challenges.

← Back to all benchmarks

Compare:
Open comparison →