Protocol

Public FEP benchmark, FEP+ accuracy against experiment

Scores relative binding free energies from a free energy perturbation workflow against measured affinities over every congeneric series in the collection, reporting compound-weighted and edge-weighted averages with bootstrap confidence intervals.

1 evaluation · 6 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, 6 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 · 6 results. Different protocols are not a single leaderboard.

Protocol: Public FEP benchmark, FEP+ accuracy against experiment · Dataset: Curated public FEP benchmark compounds (Ross et al. 2023)

Sorted by Kendall tau (of the cycle-closure-corrected absolute free energies, weighted mean over series, weight equal to 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	Kendall tau (of the cycle-closure-corrected absolute free energies, weighted mean over series, weight equal to compound count)	MAE (edgewise mean unsigned error, weighted mean over FEP graphs, weight equal to edge count)	MAE (pairwise mean unsigned error, weighted mean over series, weight equal to compound count)	R² (of the cycle-closure-corrected absolute free energies, weighted mean over series, weight equal to compound count)	RMSE (edgewise, weighted mean over FEP graphs, weight equal to edge count)	RMSE (pairwise, weighted mean over series, weight equal to compound count)
FEP+ with per-system manual protocol adjustmentEvaluation	0.51 [0.48, 0.55]	0.91 [0.84, 0.98]	0.98 [0.91, 1.05]	0.56 [0.51, 0.60]	1.17 [1.08, 1.25]	1.25 [1.17, 1.33]

All 6 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: FEP+ with per-system manual protocol adjustment	Protocol: Public FEP benchmark, FEP+ accuracy against experiment Dataset: Curated public FEP benchmark compounds (Ross et al. 2023)	0.91 [0.84, 0.98] MAE (edgewise mean unsigned error, weighted mean over FEP graphs, weight equal to edge count) kilocalorie-per-mole · lower Uncertainty: 95% CI 0.84 to 0.98. 95% confidence interval from bootstrap sampling over the pairs of experimental series or FEP+ graphs (Table 3 legend) Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source FEP+ on the curated public FEP benchmark uc-leadopt-20261022-protocol-ross-fep-plus Aggregation: Not reported The maximal and current accuracy of rigorous protein-ligand binding free energy calculations · Table 3, row "Edgewise MUE (kcal mol−1)", column "FEP+ benchmark"
Configuration: FEP+ with per-system manual protocol adjustment	Protocol: Public FEP benchmark, FEP+ accuracy against experiment Dataset: Curated public FEP benchmark compounds (Ross et al. 2023)	1.17 [1.08, 1.25] RMSE (edgewise, weighted mean over FEP graphs, weight equal to edge count) kilocalorie-per-mole · lower Uncertainty: 95% CI 1.08 to 1.25. 95% confidence interval from bootstrap sampling over the pairs of experimental series or FEP+ graphs (Table 3 legend) Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source FEP+ on the curated public FEP benchmark uc-leadopt-20261022-protocol-ross-fep-plus Aggregation: Not reported The maximal and current accuracy of rigorous protein-ligand binding free energy calculations · Table 3, row "Edgewise RMSE (kcal mol−1)", column "FEP+ benchmark"
Configuration: FEP+ with per-system manual protocol adjustment	Protocol: Public FEP benchmark, FEP+ accuracy against experiment Dataset: Curated public FEP benchmark compounds (Ross et al. 2023)	0.51 [0.48, 0.55] Kendall tau (of the cycle-closure-corrected absolute free energies, weighted mean over series, weight equal to compound count) unitless · higher Uncertainty: 95% CI 0.48 to 0.55. 95% confidence interval from bootstrap sampling over the pairs of experimental series or FEP+ graphs (Table 3 legend) Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source FEP+ on the curated public FEP benchmark uc-leadopt-20261022-protocol-ross-fep-plus Aggregation: Not reported The maximal and current accuracy of rigorous protein-ligand binding free energy calculations · Table 3, row "Kendall τ", column "FEP+ benchmark"
Configuration: FEP+ with per-system manual protocol adjustment	Protocol: Public FEP benchmark, FEP+ accuracy against experiment Dataset: Curated public FEP benchmark compounds (Ross et al. 2023)	0.98 [0.91, 1.05] MAE (pairwise mean unsigned error, weighted mean over series, weight equal to compound count) kilocalorie-per-mole · lower Uncertainty: 95% CI 0.91 to 1.05. 95% confidence interval from bootstrap sampling over the pairs of experimental series or FEP+ graphs (Table 3 legend) Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source FEP+ on the curated public FEP benchmark uc-leadopt-20261022-protocol-ross-fep-plus Aggregation: Not reported The maximal and current accuracy of rigorous protein-ligand binding free energy calculations · Table 3, row "Pairwise MUE (kcal mol−1)", column "FEP+ benchmark"
Configuration: FEP+ with per-system manual protocol adjustment	Protocol: Public FEP benchmark, FEP+ accuracy against experiment Dataset: Curated public FEP benchmark compounds (Ross et al. 2023)	1.25 [1.17, 1.33] RMSE (pairwise, weighted mean over series, weight equal to compound count) kilocalorie-per-mole · lower Uncertainty: 95% CI 1.17 to 1.33. 95% confidence interval from bootstrap sampling over the pairs of experimental series or FEP+ graphs (Table 3 legend) Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source FEP+ on the curated public FEP benchmark uc-leadopt-20261022-protocol-ross-fep-plus Aggregation: Not reported The maximal and current accuracy of rigorous protein-ligand binding free energy calculations · Table 3, row "Pairwise RMSE (kcal mol−1)", column "FEP+ benchmark"
Configuration: FEP+ with per-system manual protocol adjustment	Protocol: Public FEP benchmark, FEP+ accuracy against experiment Dataset: Curated public FEP benchmark compounds (Ross et al. 2023)	0.56 [0.51, 0.60] R² (of the cycle-closure-corrected absolute free energies, weighted mean over series, weight equal to compound count) unitless · higher Uncertainty: 95% CI 0.51 to 0.60. 95% confidence interval from bootstrap sampling over the pairs of experimental series or FEP+ graphs (Table 3 legend) Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source FEP+ on the curated public FEP benchmark uc-leadopt-20261022-protocol-ross-fep-plus Aggregation: Not reported The maximal and current accuracy of rigorous protein-ligand binding free energy calculations · Table 3, row "R2", column "FEP+ benchmark"

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.

FEP+ on the curated public FEP benchmark

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

No-change prediction under matched control conditions

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

Training-only mean-effect or linear prediction

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

The maximal and current accuracy of rigorous protein-ligand binding free energy calculations · Original source · Communications Chemistry 6:222 (2023); Europe PMC full-text XML snapshot PMC10576784

Download this release (gzip)

Technical metadata and extraction receipts

Stable ID: uc-leadopt-20261022-protocol-ross-fep-plus

areas: molecular-interactions
contexts: research
protocol: Pairwise and edgewise RMSE and MUE in kcal/mol, coefficient of determination and Kendall tau of the cycle-closure-corrected absolute free energies, each a weighted average over FEP graphs
version: Ross et al. 2023 Table 3, 'FEP+ benchmark' column
aggregation: Weighted average over FEP graphs; the weight is the number of compounds in the graph for the pairwise, R2 and Kendall tau metrics and the number of edges for the edgewise errors (Table 3 legend)
metric definition: Pairwise errors are computed over the inferred predicted free energy difference between every pair of ligands in a graph; edgewise errors over the graph's edges only and are therefore dependent on the graph topology (Methods, 'Accuracy metrics')
uncertainty reporting: definition: 95% confidence intervals from bootstrap sampling over the pairs of experimental series or FEP+ graphs (Table 3 legend)
heldout test: Not applicable: the workflow is not trained
source locator: Table 3, column 'FEP+ benchmark'
limitations: The FEP+ protocol was tuned per system after seeing the error against experiment: structural inputs, rotamer and protonation states, atom mappings and torsion scaling were adjusted by visual inspection and manual intervention when poor sampling was observed in prior simulations. The reported accuracy is maximum-attainable, not prospective.; Every value is a weighted average over series; no per-series value is printed in the paper's tables.; The authors developed the workflow being assessed.; Running this workflow needs a commercial Schrodinger licence, so its values can only ever be recorded from publication, never re-run here.; Compute cost is not reported.
missing metadata: capacity: reason: unreported; note: No compute cost is reported for the FEP+ calculations
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: f371d2e373306588d11b695b8de21b253a90293d6a15884f635bcf927bdebb1d; retrieval url: https://www.ebi.ac.uk/europepmc/webservices/rest/PMC10576784/fullTextXML

Related records

part of: Public binding free energy benchmark (Ross et al. 2023)
uses data: Curated public FEP benchmark compounds (Ross et al. 2023)
evaluates task: Small-molecule affinity
assessment: FEP+ on the curated public FEP benchmark
assessed by: Rank close analogues of a hit by binding affinity

Suggest a correction