Configuration

BACPI ligand-only ablation (GAT)

BACPI with its protein sequence network switched off, so that affinity is predicted from the ligand alone, which estimates how much of the signal is ligand bias.

3 evaluations · 28 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.

Evaluations and results

3 evaluations · 28 results. Different protocols are not a single leaderboard.

Sorted by Kendall tau (target average as printed, computed per assay and weighted by compound count) (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	Protocol and dataset	Kendall tau (target average as printed, computed per assay and weighted by compound count)	MAE (centered, computed per assay and averaged across assays weighted by compound count)	MAE (non-centered, computed per assay and averaged across assays weighted by compound count)	MAE (pairwise, target average as printed, computed per assay and weighted by compound count)	Pearson (target average as printed, computed per assay and weighted by compound count)	Proportion (within 1 kcal/mol, centered, computed per assay and averaged across assays weighted by compound count)	Proportion (within 1 kcal/mol, non-centered, computed per assay and averaged across assays weighted by compound count)	Proportion (within 2 kcal/mol, centered, computed per assay and averaged across assays weighted by compound count)	Proportion (within 2 kcal/mol, non-centered, computed per assay and averaged across assays weighted by compound count)	Accelerator time (per-method cost as printed beside the accuracy columns)
BACPI ligand-only ablation (GAT)Evaluation	CASP16 affinity challenge, Boltz-2 against the top-ranked participants, CASP16 affinity targets L1000 and L3000	0.35 (best in column)	1.13	1.28	1.58	0.5 (best in column)	0.49	0.44 (best in column)	0.84	0.79	Not reported
BACPI ligand-only ablation (GAT)Evaluation	FEP+ benchmark four-target subset, affinity ranking and cost, Four-target subset of the public FEP benchmark	0.28	0.71 (best in column)	1.19 (best in column)	1.07 (best in column)	0.4	0.72 (best in column)	0.43	0.95 (best in column)	0.86 (best in column)	0.18 GPU ms
BACPI ligand-only ablation (GAT)Evaluation	FEP+ benchmark OpenFE subset, affinity ranking, OpenFE subset of the public FEP benchmark	0.2	0.91	1.42	1.3	0.28	0.64	0.4	0.92	0.75	Not reported

All 28 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: BACPI ligand-only ablation (GAT)	Protocol: FEP+ benchmark OpenFE subset, affinity ranking Dataset: OpenFE subset of the public FEP benchmark	0.2 Kendall tau (target average as printed, computed per assay and weighted by compound count) unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source GAT on the FEP+ benchmark OpenFE subset uc-leadopt-20261022-protocol-boltz2-fep-openfe-subset Aggregation: Not reported Boltz-2: Towards Accurate and Efficient Binding Affinity Prediction, supplementary appendix · Table 11, row "GAT", column "Kendall tau", sub-column "target avg." (text-layer line 1692)
Configuration: BACPI ligand-only ablation (GAT)	Protocol: FEP+ benchmark OpenFE subset, affinity ranking Dataset: OpenFE subset of the public FEP benchmark	0.91 MAE (centered, computed per assay and averaged across assays weighted by compound count) kilocalorie-per-mole · lower Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source GAT on the FEP+ benchmark OpenFE subset uc-leadopt-20261022-protocol-boltz2-fep-openfe-subset Aggregation: Not reported Boltz-2: Towards Accurate and Efficient Binding Affinity Prediction, supplementary appendix · Table 11, row "GAT", column "MAE", sub-column "cent." (text-layer line 1692)
Configuration: BACPI ligand-only ablation (GAT)	Protocol: FEP+ benchmark OpenFE subset, affinity ranking Dataset: OpenFE subset of the public FEP benchmark	1.42 MAE (non-centered, computed per assay and averaged across assays weighted by compound count) kilocalorie-per-mole · lower Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source GAT on the FEP+ benchmark OpenFE subset uc-leadopt-20261022-protocol-boltz2-fep-openfe-subset Aggregation: Not reported Boltz-2: Towards Accurate and Efficient Binding Affinity Prediction, supplementary appendix · Table 11, row "GAT", column "MAE", sub-column "non-cent." (text-layer line 1692)
Configuration: BACPI ligand-only ablation (GAT)	Protocol: FEP+ benchmark OpenFE subset, affinity ranking Dataset: OpenFE subset of the public FEP benchmark	0.28 Pearson (target average as printed, computed per assay and weighted by compound count) unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source GAT on the FEP+ benchmark OpenFE subset uc-leadopt-20261022-protocol-boltz2-fep-openfe-subset Aggregation: Not reported Boltz-2: Towards Accurate and Efficient Binding Affinity Prediction, supplementary appendix · Table 11, row "GAT", column "Pearson R", sub-column "target avg." (text-layer line 1692)
Configuration: BACPI ligand-only ablation (GAT)	Protocol: FEP+ benchmark OpenFE subset, affinity ranking Dataset: OpenFE subset of the public FEP benchmark	1.3 MAE (pairwise, target average as printed, computed per assay and weighted by compound count) kilocalorie-per-mole · lower Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source GAT on the FEP+ benchmark OpenFE subset uc-leadopt-20261022-protocol-boltz2-fep-openfe-subset Aggregation: Not reported Boltz-2: Towards Accurate and Efficient Binding Affinity Prediction, supplementary appendix · Table 11, row "GAT", column "PMAE", sub-column "target avg." (text-layer line 1692)
Configuration: BACPI ligand-only ablation (GAT)	Protocol: FEP+ benchmark OpenFE subset, affinity ranking Dataset: OpenFE subset of the public FEP benchmark	0.64 Proportion (within 1 kcal/mol, centered, computed per assay and averaged across assays weighted by compound count) fraction · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source GAT on the FEP+ benchmark OpenFE subset uc-leadopt-20261022-protocol-boltz2-fep-openfe-subset Aggregation: Not reported Boltz-2: Towards Accurate and Efficient Binding Affinity Prediction, supplementary appendix · Table 11, row "GAT", column "Perc. within 1", sub-column "cent." (text-layer line 1692)
Configuration: BACPI ligand-only ablation (GAT)	Protocol: FEP+ benchmark OpenFE subset, affinity ranking Dataset: OpenFE subset of the public FEP benchmark	0.4 Proportion (within 1 kcal/mol, non-centered, computed per assay and averaged across assays weighted by compound count) fraction · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source GAT on the FEP+ benchmark OpenFE subset uc-leadopt-20261022-protocol-boltz2-fep-openfe-subset Aggregation: Not reported Boltz-2: Towards Accurate and Efficient Binding Affinity Prediction, supplementary appendix · Table 11, row "GAT", column "Perc. within 1", sub-column "non-cent." (text-layer line 1692)
Configuration: BACPI ligand-only ablation (GAT)	Protocol: FEP+ benchmark OpenFE subset, affinity ranking Dataset: OpenFE subset of the public FEP benchmark	0.92 Proportion (within 2 kcal/mol, centered, computed per assay and averaged across assays weighted by compound count) fraction · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source GAT on the FEP+ benchmark OpenFE subset uc-leadopt-20261022-protocol-boltz2-fep-openfe-subset Aggregation: Not reported Boltz-2: Towards Accurate and Efficient Binding Affinity Prediction, supplementary appendix · Table 11, row "GAT", column "Perc. within 2", sub-column "cent." (text-layer line 1692)
Configuration: BACPI ligand-only ablation (GAT)	Protocol: FEP+ benchmark OpenFE subset, affinity ranking Dataset: OpenFE subset of the public FEP benchmark	0.75 Proportion (within 2 kcal/mol, non-centered, computed per assay and averaged across assays weighted by compound count) fraction · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source GAT on the FEP+ benchmark OpenFE subset uc-leadopt-20261022-protocol-boltz2-fep-openfe-subset Aggregation: Not reported Boltz-2: Towards Accurate and Efficient Binding Affinity Prediction, supplementary appendix · Table 11, row "GAT", column "Perc. within 2", sub-column "non-cent." (text-layer line 1692)
Configuration: BACPI ligand-only ablation (GAT)	Protocol: FEP+ benchmark four-target subset, affinity ranking and cost Dataset: Four-target subset of the public FEP benchmark	0.28 Kendall tau (target average as printed, computed per assay and weighted by compound count) unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source GAT on the FEP+ benchmark four-target subset uc-leadopt-20261022-protocol-boltz2-fep-4target Aggregation: Not reported Boltz-2: Towards Accurate and Efficient Binding Affinity Prediction, supplementary appendix · Table 12, row "GAT", column "Kendall tau", sub-column "target avg." (text-layer line 1709)
Configuration: BACPI ligand-only ablation (GAT)	Protocol: FEP+ benchmark four-target subset, affinity ranking and cost Dataset: Four-target subset of the public FEP benchmark	0.71 MAE (centered, computed per assay and averaged across assays weighted by compound count) kilocalorie-per-mole · lower Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source GAT on the FEP+ benchmark four-target subset uc-leadopt-20261022-protocol-boltz2-fep-4target Aggregation: Not reported Boltz-2: Towards Accurate and Efficient Binding Affinity Prediction, supplementary appendix · Table 12, row "GAT", column "MAE", sub-column "cent." (text-layer line 1709)
Configuration: BACPI ligand-only ablation (GAT)	Protocol: FEP+ benchmark four-target subset, affinity ranking and cost Dataset: Four-target subset of the public FEP benchmark	1.19 MAE (non-centered, computed per assay and averaged across assays weighted by compound count) kilocalorie-per-mole · lower Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source GAT on the FEP+ benchmark four-target subset uc-leadopt-20261022-protocol-boltz2-fep-4target Aggregation: Not reported Boltz-2: Towards Accurate and Efficient Binding Affinity Prediction, supplementary appendix · Table 12, row "GAT", column "MAE", sub-column "non-cent." (text-layer line 1709)
Configuration: BACPI ligand-only ablation (GAT)	Protocol: FEP+ benchmark four-target subset, affinity ranking and cost Dataset: Four-target subset of the public FEP benchmark	0.4 Pearson (target average as printed, computed per assay and weighted by compound count) unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source GAT on the FEP+ benchmark four-target subset uc-leadopt-20261022-protocol-boltz2-fep-4target Aggregation: Not reported Boltz-2: Towards Accurate and Efficient Binding Affinity Prediction, supplementary appendix · Table 12, row "GAT", column "Pearson R", sub-column "target avg." (text-layer line 1709)
Configuration: BACPI ligand-only ablation (GAT)	Protocol: FEP+ benchmark four-target subset, affinity ranking and cost Dataset: Four-target subset of the public FEP benchmark	1.07 MAE (pairwise, target average as printed, computed per assay and weighted by compound count) kilocalorie-per-mole · lower Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source GAT on the FEP+ benchmark four-target subset uc-leadopt-20261022-protocol-boltz2-fep-4target Aggregation: Not reported Boltz-2: Towards Accurate and Efficient Binding Affinity Prediction, supplementary appendix · Table 12, row "GAT", column "PMAE", sub-column "target avg." (text-layer line 1709)
Configuration: BACPI ligand-only ablation (GAT)	Protocol: FEP+ benchmark four-target subset, affinity ranking and cost Dataset: Four-target subset of the public FEP benchmark	0.18 GPU ms Accelerator time (per-method cost as printed beside the accuracy columns) millisecond · lower Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source GAT on the FEP+ benchmark four-target subset uc-leadopt-20261022-protocol-boltz2-fep-4target Aggregation: Not reported Boltz-2: Towards Accurate and Efficient Binding Affinity Prediction, supplementary appendix · Table 12, row "GAT", column "time" (text-layer line 1709)
Configuration: BACPI ligand-only ablation (GAT)	Protocol: FEP+ benchmark four-target subset, affinity ranking and cost Dataset: Four-target subset of the public FEP benchmark	0.72 Proportion (within 1 kcal/mol, centered, computed per assay and averaged across assays weighted by compound count) fraction · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source GAT on the FEP+ benchmark four-target subset uc-leadopt-20261022-protocol-boltz2-fep-4target Aggregation: Not reported Boltz-2: Towards Accurate and Efficient Binding Affinity Prediction, supplementary appendix · Table 12, row "GAT", column "Perc. within 1", sub-column "cent." (text-layer line 1709)
Configuration: BACPI ligand-only ablation (GAT)	Protocol: FEP+ benchmark four-target subset, affinity ranking and cost Dataset: Four-target subset of the public FEP benchmark	0.43 Proportion (within 1 kcal/mol, non-centered, computed per assay and averaged across assays weighted by compound count) fraction · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source GAT on the FEP+ benchmark four-target subset uc-leadopt-20261022-protocol-boltz2-fep-4target Aggregation: Not reported Boltz-2: Towards Accurate and Efficient Binding Affinity Prediction, supplementary appendix · Table 12, row "GAT", column "Perc. within 1", sub-column "non-cent." (text-layer line 1709)
Configuration: BACPI ligand-only ablation (GAT)	Protocol: FEP+ benchmark four-target subset, affinity ranking and cost Dataset: Four-target subset of the public FEP benchmark	0.95 Proportion (within 2 kcal/mol, centered, computed per assay and averaged across assays weighted by compound count) fraction · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source GAT on the FEP+ benchmark four-target subset uc-leadopt-20261022-protocol-boltz2-fep-4target Aggregation: Not reported Boltz-2: Towards Accurate and Efficient Binding Affinity Prediction, supplementary appendix · Table 12, row "GAT", column "Perc. within 2", sub-column "cent." (text-layer line 1709)
Configuration: BACPI ligand-only ablation (GAT)	Protocol: FEP+ benchmark four-target subset, affinity ranking and cost Dataset: Four-target subset of the public FEP benchmark	0.86 Proportion (within 2 kcal/mol, non-centered, computed per assay and averaged across assays weighted by compound count) fraction · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source GAT on the FEP+ benchmark four-target subset uc-leadopt-20261022-protocol-boltz2-fep-4target Aggregation: Not reported Boltz-2: Towards Accurate and Efficient Binding Affinity Prediction, supplementary appendix · Table 12, row "GAT", column "Perc. within 2", sub-column "non-cent." (text-layer line 1709)
Configuration: BACPI ligand-only ablation (GAT)	Protocol: CASP16 affinity challenge, Boltz-2 against the top-ranked participants Dataset: CASP16 affinity targets L1000 and L3000	0.35 Kendall tau (target average as printed, computed per assay and weighted by compound count) unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source GAT on the CASP16 affinity challenge uc-leadopt-20261022-protocol-boltz2-casp16 Aggregation: Not reported Boltz-2: Towards Accurate and Efficient Binding Affinity Prediction, supplementary appendix · Table 14, row "GAT", column "Kendall tau", sub-column "target avg." (text-layer line 1755)
Configuration: BACPI ligand-only ablation (GAT)	Protocol: CASP16 affinity challenge, Boltz-2 against the top-ranked participants Dataset: CASP16 affinity targets L1000 and L3000	1.13 MAE (centered, computed per assay and averaged across assays weighted by compound count) kilocalorie-per-mole · lower Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source GAT on the CASP16 affinity challenge uc-leadopt-20261022-protocol-boltz2-casp16 Aggregation: Not reported Boltz-2: Towards Accurate and Efficient Binding Affinity Prediction, supplementary appendix · Table 14, row "GAT", column "MAE", sub-column "cent." (text-layer line 1755)
Configuration: BACPI ligand-only ablation (GAT)	Protocol: CASP16 affinity challenge, Boltz-2 against the top-ranked participants Dataset: CASP16 affinity targets L1000 and L3000	1.28 MAE (non-centered, computed per assay and averaged across assays weighted by compound count) kilocalorie-per-mole · lower Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source GAT on the CASP16 affinity challenge uc-leadopt-20261022-protocol-boltz2-casp16 Aggregation: Not reported Boltz-2: Towards Accurate and Efficient Binding Affinity Prediction, supplementary appendix · Table 14, row "GAT", column "MAE", sub-column "non-cent." (text-layer line 1755)
Configuration: BACPI ligand-only ablation (GAT)	Protocol: CASP16 affinity challenge, Boltz-2 against the top-ranked participants Dataset: CASP16 affinity targets L1000 and L3000	0.5 Pearson (target average as printed, computed per assay and weighted by compound count) unitless · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source GAT on the CASP16 affinity challenge uc-leadopt-20261022-protocol-boltz2-casp16 Aggregation: Not reported Boltz-2: Towards Accurate and Efficient Binding Affinity Prediction, supplementary appendix · Table 14, row "GAT", column "Pearson R", sub-column "target avg." (text-layer line 1755)
Configuration: BACPI ligand-only ablation (GAT)	Protocol: CASP16 affinity challenge, Boltz-2 against the top-ranked participants Dataset: CASP16 affinity targets L1000 and L3000	1.58 MAE (pairwise, target average as printed, computed per assay and weighted by compound count) kilocalorie-per-mole · lower Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source GAT on the CASP16 affinity challenge uc-leadopt-20261022-protocol-boltz2-casp16 Aggregation: Not reported Boltz-2: Towards Accurate and Efficient Binding Affinity Prediction, supplementary appendix · Table 14, row "GAT", column "Pairwise MAE", sub-column "target avg." (text-layer line 1755)
Configuration: BACPI ligand-only ablation (GAT)	Protocol: CASP16 affinity challenge, Boltz-2 against the top-ranked participants Dataset: CASP16 affinity targets L1000 and L3000	0.49 Proportion (within 1 kcal/mol, centered, computed per assay and averaged across assays weighted by compound count) fraction · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source GAT on the CASP16 affinity challenge uc-leadopt-20261022-protocol-boltz2-casp16 Aggregation: Not reported Boltz-2: Towards Accurate and Efficient Binding Affinity Prediction, supplementary appendix · Table 14, row "GAT", column "Perc. within 1", sub-column "cent." (text-layer line 1755)

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

Use this configuration

How it works, versions and access

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

Boltz-2: Towards Accurate and Efficient Binding Affinity Prediction, supplementary appendix · Original source · bioRxiv v1 posted 2025-06-18; supplement PDF as deposited in Europe PMC PMC12262699 (PPR1038344)

Download this release (gzip)

Technical metadata and extraction receipts

Stable ID: uc-leadopt-20261022-config-gat-ligand-only

areas: molecular-interactions
contexts: research
method types: supervised_machine_learning
configuration type: paper_evaluation
reported name: GAT
version: Retrained for the Boltz-2 comparison; no release version printed
hyperparameters: Default hyperparameters of the original publication except batch size, raised to 32; the protein sequence CNN is deactivated
access reuse: Not stated in this source; see the BACPI release.
source locator: Tables 11, 12 and 14, row 'GAT'; appendix D.2.2
review: method: source-hash-verification; pdf-text-parse; independent-cell-check; rendered-source-visual-check; ai-assisted-source-review; method note: Re-retrieved all three sources and matched each pinned SHA-256. Re-rendered the Boltz-2 supplement with pdftotext -layout (poppler 26.08.0) and matched the text layer's recorded SHA-256. Re-read all 256 cells with a parser written for this review that does not import the extraction script, matching each result to its cell through its own source_locator, and separately read appendix tables 11, 12 and 14 from PNG renderings of PDF pages 29 and 30 at 160 dpi, and Ross Table 3 from the pinned XML markup. Checked printed value, numeric value, locator, metric concept, qualifier, direction, unit, uncertainty and the linked configuration, protocol and dataset, and that every N/A or '-' cell is recorded as missing with a reason.; reviewer: claude; reviewer note: Claude review agent, independent of the extraction agent; no human review claimed; reviewed at: 2026-10-22T13:40:00Z; artifact sha256: be867a5d92add8b223d47d882a4213ad489741f3e4d5081a0901a54370342569; retrieval url: https://www.ebi.ac.uk/europepmc/webservices/rest/PMC12262699/supplementaryFiles

Related records

configuration of: BACPI
system: GAT on the FEP+ benchmark OpenFE subset
system: GAT on the FEP+ benchmark four-target subset
system: GAT on the CASP16 affinity challenge

Suggest a correction