Protocol

BindCraft experimental success rate averaged over its 12 benchmark targets

Evidence concern: excluded from comparisons

Supplementary Data 1 has one row per tested design with a Binding column holding 1, 0 or an empty cell. Treating the non-empty cells as the tested set reproduces the prose counts for BBF-14 (11 tested, 6 binding), Bet v1 (7, 2), CD45 (16, 4), Der f7 (10, 4), Der f21 (7, 4), IFNAR2 (9, 3), PD-1 (53, 13), SpCas9 (6, 6) and CLDN1 (7, 6), but not for PD-L1 (13 non-empty rows with 11 binding against 9 tested with 7 binding in the prose), CrSAS-6 (10 non-empty rows against 9 tested) or CbAgo (2 non-empty rows against 12 tested). The row filter is therefore ambiguous and no value is computed from these rows.One-shot design of functional protein binders with BindCraft (Pacesa et al., Nature 2025) · Supplementary Data 1 (41586_2025_9429_MOESM4_ESM.csv), columns Target, DesignName, Binding; Results prose; Fig. 1b caption

Fig. 1b's caption defines success as binding observed on SPR, but the PD-1 screen was BLI in a bivalent Fc-fusion format and the CLDN1 screen used soluble analogues. The 46.3% average therefore pools targets measured under different endpoints.One-shot design of functional protein binders with BindCraft (Pacesa et al., Nature 2025) · Fig. 1b caption; Results 'Binders targeting cell-surface receptors' and 'Targeting unexplored binding sites'

Mean and range of the per-target experimental success rate over the 12 targets BindCraft was benchmarked on.

1 evaluation · 2 results

Overview

An explanatory profile has not yet been reviewed for this record.

Consult the linked sources for architecture or protocol details. Missing evidence is not evidence of a missing capability.

1 recorded evaluation, 2 metric rows. A comparison chart has not yet been validated for these results. The table retains the individual findings and their sources.

View coverage and remaining gaps across all benchmarks

Results

Results are available, but no reviewed comparison panel is linked in this release.

All evaluations

1 evaluation · 2 results. Different protocols are not a single leaderboard.

Protocol: BindCraft experimental success rate averaged over its 12 benchmark targets · Dataset: BindCraft 12-target benchmark set

Sorted by Hit rate (range of per-target experimental success rates over 12 targets) (higher is better). The best value in each column is highlighted. Decimals are rounded for display; each value links to the printed value and its source.

Tested configuration	Hit rate (range of per-target experimental success rates over 12 targets)	Hit rate (unweighted mean of per-target experimental success rates over 12 targets)
BindCraft default settingsEvaluation	10% to 100%	46.3%

All 2 result rows with coverage, uncertainty and sources

Filter evaluations

Tested model or methodProtocolDatasetEvaluation setupMetricEvidence origin

Applied filters: All linked evaluations

Exact evaluated configurations and original reported results
Tested configuration	Protocol and dataset	Finding	Evidence and details
Configuration: BindCraft default settings	Protocol: BindCraft experimental success rate averaged over its 12 benchmark targets Dataset: BindCraft 12-target benchmark set	46.3% Hit rate (unweighted mean of per-target experimental success rates over 12 targets) percent · higher Uncertainty: Not reported by the source: No interval is printed. Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Evidence concern: excluded from comparisons Supplementary Data 1 has one row per tested design with a Binding column holding 1, 0 or an empty cell. Treating the non-empty cells as the tested set reproduces the prose counts for BBF-14 (11 tested, 6 binding), Bet v1 (7, 2), CD45 (16, 4), Der f7 (10, 4), Der f21 (7, 4), IFNAR2 (9, 3), PD-1 (53, 13), SpCas9 (6, 6) and CLDN1 (7, 6), but not for PD-L1 (13 non-empty rows with 11 binding against 9 tested with 7 binding in the prose), CrSAS-6 (10 non-empty rows against 9 tested) or CbAgo (2 non-empty rows against 12 tested). The row filter is therefore ambiguous and no value is computed from these rows.One-shot design of functional protein binders with BindCraft (Pacesa et al., Nature 2025) · Supplementary Data 1 (41586_2025_9429_MOESM4_ESM.csv), columns Target, DesignName, Binding; Results prose; Fig. 1b caption Fig. 1b's caption defines success as binding observed on SPR, but the PD-1 screen was BLI in a bivalent Fc-fusion format and the CLDN1 screen used soluble analogues. The 46.3% average therefore pools targets measured under different endpoints.One-shot design of functional protein binders with BindCraft (Pacesa et al., Nature 2025) · Fig. 1b caption; Results 'Binders targeting cell-surface receptors' and 'Targeting unexplored binding sites' Methods, coverage and source BindCraft experimental success rate over its 12 benchmark targets uc-binder-20261022-protocol-bindcraft-mean-success-rate-12-targets Aggregation: Not reported One-shot design of functional protein binders with BindCraft (Pacesa et al., Nature 2025) · Introduction, final paragraph: 'The binder affinities lie predominantly in the nanomolar range, with success rates from 10% to 100%, with an average success rate of 46.3%'
Configuration: BindCraft default settings	Protocol: BindCraft experimental success rate averaged over its 12 benchmark targets Dataset: BindCraft 12-target benchmark set	10% to 100% Hit rate (range of per-target experimental success rates over 12 targets) percent · higher Uncertainty: Not applicable: The value is itself a range over targets, not an estimate with an interval. Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Evidence concern: excluded from comparisons Supplementary Data 1 has one row per tested design with a Binding column holding 1, 0 or an empty cell. Treating the non-empty cells as the tested set reproduces the prose counts for BBF-14 (11 tested, 6 binding), Bet v1 (7, 2), CD45 (16, 4), Der f7 (10, 4), Der f21 (7, 4), IFNAR2 (9, 3), PD-1 (53, 13), SpCas9 (6, 6) and CLDN1 (7, 6), but not for PD-L1 (13 non-empty rows with 11 binding against 9 tested with 7 binding in the prose), CrSAS-6 (10 non-empty rows against 9 tested) or CbAgo (2 non-empty rows against 12 tested). The row filter is therefore ambiguous and no value is computed from these rows.One-shot design of functional protein binders with BindCraft (Pacesa et al., Nature 2025) · Supplementary Data 1 (41586_2025_9429_MOESM4_ESM.csv), columns Target, DesignName, Binding; Results prose; Fig. 1b caption Fig. 1b's caption defines success as binding observed on SPR, but the PD-1 screen was BLI in a bivalent Fc-fusion format and the CLDN1 screen used soluble analogues. The 46.3% average therefore pools targets measured under different endpoints.One-shot design of functional protein binders with BindCraft (Pacesa et al., Nature 2025) · Fig. 1b caption; Results 'Binders targeting cell-surface receptors' and 'Targeting unexplored binding sites' Methods, coverage and source BindCraft experimental success rate over its 12 benchmark targets uc-binder-20261022-protocol-bindcraft-mean-success-rate-12-targets Aggregation: Not reported One-shot design of functional protein binders with BindCraft (Pacesa et al., Nature 2025) · Introduction, final paragraph: 'with success rates from 10% to 100%'

Source checking is not independent reproduction. Release 2026-10-22-1b7cf027e4f6.

Methods and evaluation design

Procedure, tasks and evaluated configurations

Recorded evaluations

Each evaluation records what was tested and under which conditions.

BindCraft experimental success rate over its 12 benchmark targets

Baseline coverage

Reference methods help show what a model adds beyond simple controls. We track a null control and a conventional method for each protocol.

0 of 2 active baseline roles have published Rewire measurements in this release. Measurements on a selected protocol do not establish coverage of an entire suite.

No execution recipe linked to this protocol. Recipe availability does not establish a completed evaluation.

Author-reported evaluations: 1

Literature evidence is not a Rewire measurement. Executed but unpublished runs and private review status are not included.

Null control

Proposed control: requires review

Select a task-valid null control after reviewing inputs and metric

Protocol-specific applicability, permitted inputs, access, split, evaluator and execution requirements need review before implementation or execution.

This is a suggested selection rule, not a validated method or a measured score.

Conventional reference

Proposed control: requires review

Select an upstream conventional reference after reviewing the full protocol

Protocol-specific applicability, permitted inputs, access, split, evaluator and execution requirements need review before implementation or execution.

This is a suggested selection rule, not a validated method or a measured score.

Protocol coverage CSV (gzip) · Model evaluation matrix (gzip) · Source table (gzip) · Release and checksums (gzip)

Coverage is derived from release 2026-10-22-1b7cf027e4f6. Source citations describe the original records; they do not validate an unreviewed baseline proposal. No results have been generated by this audit.

Run instructions

No runnable recipe has been reviewed for this protocol. Dataset access, model requirements, licences and compute requirements must be checked against its sources before execution.

Strengths, limitations and unresolved questions

Evidence

Source checking verifies the cited claim or transcription. It does not establish independent reproduction.

Evidence table

Inspect claims, sources and review details

Trace each statement to its source and review. A context-only reference supports the record generally; it does not verify an individual field. Source checking does not reproduce an experiment.

One row per statement and cited source. Multiple citations are not independent evaluations. Shared locators are labelled explicitly.

0 evidence rows matching the loaded filters

Claims, original sources and review scope · Release 2026-10-22-1b7cf027e4f6
Property and statement	Original source and location	Review and provenance

No evidence rows match these filters. Choose another scope or clear the search.

How to read the evidence

Sources and history

Release 2026-10-22-1b7cf027e4f6 · Record review: source checked

1 source record and release history

One-shot design of functional protein binders with BindCraft (Pacesa et al., Nature 2025) · Original source · Nature, version of record; Europe PMC JATS full text for PMC12507698

Download this release (gzip)

Technical metadata and extraction receipts

Stable ID: uc-binder-20261022-protocol-bindcraft-mean-success-rate-12-targets

areas: proteins-complexes; protein-structure
contexts: research
aggregation: Unweighted mean over the 12 benchmark targets; each target contributes its own rate regardless of how many designs were tested.
comparison limit: An unweighted mean over targets whose denominators range from 6 to 53 designs, and whose endpoints are not the same assay. It is not a hit rate over a pooled design set.
denominator: 12
limitations: Developer-reported: the BindCraft authors built the pipeline, chose the targets and chose which designs to test.; The assay is not uniform across targets. Fig. 1b's caption says binding was observed on SPR, but PD-1 was screened by BLI in a bivalent Fc-fusion format and the claudin binders were prescreened against soluble analogues of the target rather than the wild-type protein.; No numeric binding threshold is stated. A successful design is one where a binding signal was observed.; Denominators are 6 to 53 designs per target, and the designs tested are about 10 of the top 20 ranked by the pipeline's own i_pTM filter, so design and selection are coupled.; No conventional comparator (directed evolution, display selection or grafting) was run on these targets.; No uncertainty is printed.; Two targets, CrSAS-6 and CbAgo, have a stated denominator but no stated numerator in the prose; their numerators appear only inside the Fig. 1b graphic, which was not read.; An unweighted mean over targets, not a pooled rate over designs. A target where 6 of 6 designs bound contributes as much as one where 13 of 53 did.; The 12 targets were chosen by the authors.; The source does not state how the average was computed. The string '46.3' appears once in the article, as 'an average success rate of 46.3%', with no definition of the basis; the `aggregation` and `comparison_limit` fields on this protocol record the extraction pass's reading of it as an unweighted mean over the twelve targets, not a basis the source states. Neither candidate basis reconciles with the printed per-target counts: an unweighted mean over the twelve targets admits no integer numerators for CrSAS-6 (of 9) and CbAgo (of 12) that yield a displayed 46.3%, and neither does a pooled rate over all 156 designs. The arithmetic is in the source_warnings on uc-binder-20261022-result-bindcraft-mean-success-rate.
metric: hit-rate
metric direction: higher
metric qualifier: mean of per-target experimental success rates over 12 targets
protocol: Per-target experimental success rates from the Fig. 1b screens, averaged over the 12 benchmark targets and reported with their range.
review: artifact sha256: 74bd2902d46460ec387ed6ee46b5bf8fe61b9ea09f2d7be4daa2046ac7e59d5d; method: source-hash-verification; independent-cell-check; ai-assisted-source-review; method note: Re-downloaded from Europe PMC (HTTP 200, 226,059 bytes) and the SHA-256 recomputed and matched. Own flattening of the JATS XML; every quoted locator sentence and every printed value checked verbatim against it by a parser written for this review.; note: Reviewed in docs/reviews/use-cases/binder-2026-10-22.md. Automated AI review of the source transcription only; it is not human scientific review and nothing here was reproduced experimentally.; retrieval url: https://www.ebi.ac.uk/europepmc/webservices/rest/PMC12507698/fullTextXML; reviewed at: 2026-10-22T15:00:00Z; reviewer: claude; reviewer note: Claude review agent, independent of the extraction agent; no human review claimed
source locator: Introduction, final paragraph
unit: percent
version: Nature 2025 version of record; Introduction

Related records

assessment: BindCraft experimental success rate over its 12 benchmark targets
assessed by: Design a new binder to a chosen target

Suggest a correction