NetGoodIndexSubmit a correction

Benefit Ledger · verified · Biology

BindCraft designs nanomolar binders without high-throughput screening

EPFL-led BindCraft, published in Nature on 27 August 2025, used AlphaFold2 weights to design binders with 10–100 percent experimental success across diverse targets, including allergen, Cas9, toxin, and AAV applications.

27 Aug 2025Tier 3 MajorMethodology 0.1

Current score

+1.40

10 base · Major (tier 3 of 5, 10 pts)
× 0.5500 attribution · Critical contribution
× 0.7000 evidence · Peer review or independent validation
× 0.6000 realization · Experimentally validated
× 0.7000 durability
Event-level product before credit split: 1.62

Major practical binder method (tier 3). Attribution split with AlphaFold2 (0.55). Nature plus open code. Realization includes functional assays beyond binding. Durable pipeline.

What happened

The open pipeline reported one-shot binder design without known sites in some cases and showed functional effects in patient-derived IgE samples, Cas9 modulation, toxin cytotoxicity reduction, and AAV retargeting. It depends on AlphaFold2 as a scoring/design prior, so credit is shared with that model family.

Model attribution

BindCraft

Open pipeline that optimized binders using AF2-based losses.

Pipeline and experimental campaign are new; AF2 weights are essential.

Attribution 0.5500 · Credit share 55.00000000000001% · EPFL

AlphaFold

Provided weights used as the design/scoring engine.

BindCraft explicitly leverages AF2 weights; AF2 was not trained for this campaign.

Attribution 0.5000 · Credit share 35% · Google DeepMind

Claims

  • BindCraft produced experimentally successful binders across multiple target classes with reported functional follow-ups.

    outcome · supported

Sources

primary sources

Secondary domains: Medicine & Health

Revision history

  • 13 Sep 2026 · 0.00 1.40

    Initial adjudicated seed score under methodology 0.1.

BindCraft designs nanomolar binders without high-throughput screening · NetGoodIndex