Result

Pearson (PLABench chembl35) of FLOWR.ROOT (CASP16 and ChEMBL35 evaluation, PLABench) on PLABench ChEMBL35 leakage-controlled test set

0.392 Pearson (PLABench chembl35)

Methods

Tested configuration: FLOWR.ROOT (CASP16 and ChEMBL35 evaluation, PLABench)
Protocol: PLABench ChEMBL35 leakage-controlled affinity
Dataset: PLABench ChEMBL35 leakage-controlled test set
Procedure: PLABench ChEMBL35 leakage-controlled affinity
Evaluation: FLOWR.ROOT on PLABench chembl35
Coverage: Not reported scored / Not reported eligible
Uncertainty: Not reported by the source
Evidence: Independent external evaluation · source checkedLeakage-controlled benchmarking reveals generalization limits of deep learning for protein-ligand binding affinity prediction · Table 4, row 'FLOWR.ROOT', column 'Pearson' (text-layer line 1034)

A source-checked result verifies the numerical transcription, not every model or protocol detail. Evaluation metadata: source checked. Source checked does not mean independently reproduced.

Reproduction

Split: chembl35
Adaptation: Published model applied without retraining
Scoring implementation: As defined in Methods, 'Evaluation metrics'

No execution recipe has been verified for this exact configuration and evaluation. A benchmark's general instructions may use different inputs, splits or model settings.

Reproducing this published result requires matching its model configuration, data, split and scorer. Source checking or a successful smoke test does not establish score reproduction.

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.

1 evidence row matching the loaded filters

Claims, original sources and review scope · Release 2026-10-23-5347695500ed
Property and statement	Original source and location	Review and provenance
Reported result 0.392 Individual claims	Leakage-controlled benchmarking reveals generalization limits of deep learning for protein-ligand binding affinity prediction Original source ↗ Table 4, row 'FLOWR.ROOT', column 'Pearson' (text-layer line 1034) Version: bioRxiv 10.64898/2026.09.17.752433v1 (posted 2026-09-23); 30-page PDF Retrieved: 2026-10-11T21:28:04Z	source checked ["source-hash-verification","pdf-text-parse","independent-cell-check","rendered-source-visual-check"] · 2026-10-11 independent paper Audit details Extraction: Parsed the pdftotext -layout text layer of the pinned bioRxiv PDF with extract/extract_leakctl.py; row labels and the number of numeric cells per row asserted. Table 4, text-layer line 1034. Independent review 2026-10-11: Re-downloaded the bioRxiv PDF: its hash differs because of the per-request banner, but the pdftotext -layout text layer hashes exactly to the recorded value, so the content checked is the content extracted. Parsed all 279 cells (Table 2: 108, Table 3: 108, Table 4: 63) with an independent parser, mapping every result through its evaluation, configuration and protocol to a model, split and metric cell: all 279 printed and numeric values and all locator line numbers match, with no cell missed. Read 44 sampled values covering all three tables and all nine models from rendered pages 6, 10 and 12: 44 of 44 match. claims.csv rows, printed values and locators equal the batch. Field: `attributes.printed_value` Source artifact SHA-256: `5fba9691f7d0777d19dabc66144e5c23d40e430de4919f97b92e4d50fb397281` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact Extraction artifact SHA-256: `5fba9691f7d0777d19dabc66144e5c23d40e430de4919f97b92e4d50fb397281` Extraction artifact

How to read the evidence

Sources and history

Release 2026-10-23-5347695500ed · Record review: source checked

1 source record and release history

Leakage-controlled benchmarking reveals generalization limits of deep learning for protein-ligand binding affinity prediction · Original source · bioRxiv 10.64898/2026.09.17.752433v1 (posted 2026-09-23); 30-page PDF

Download this release (gzip)

Technical metadata and extraction receipts

Stable ID: lit-interactions-20261011-leakctl-result-flowr-root-chembl35-pearson

metric: pearson-correlation
metric direction: higher
unit: unitless
printed value: 0.392
source locator: Table 4, row 'FLOWR.ROOT', column 'Pearson' (text-layer line 1034)
review: method: source-hash-verification; pdf-text-parse; independent-cell-check; rendered-source-visual-check; method note: Re-downloaded the bioRxiv PDF: its hash differs because of the per-request banner, but the pdftotext -layout text layer hashes exactly to the recorded value, so the content checked is the content extracted. Parsed all 279 cells (Table 2: 108, Table 3: 108, Table 4: 63) with an independent parser, mapping every result through its evaluation, configuration and protocol to a model, split and metric cell: all 279 printed and numeric values and all locator line numbers match, with no cell missed. Read 44 sampled values covering all three tables and all nine models from rendered pages 6, 10 and 12: 44 of 44 match. claims.csv rows, printed values and locators equal the batch.; reviewer: claude; reviewer note: Claude review agent reviewer-mi-2, independent of the extractor lit-microbial-interactions; no human review claimed; date: 2026-10-11; artifact sha256: 5fba9691f7d0777d19dabc66144e5c23d40e430de4919f97b92e4d50fb397281; retrieval url: https://www.biorxiv.org/content/10.64898/2026.09.17.752433v1.full.pdf; note: Extraction: Parsed the pdftotext -layout text layer of the pinned bioRxiv PDF with extract/extract_leakctl.py; row labels and the number of numeric cells per row asserted. Table 4, text-layer line 1034. Independent review 2026-10-11: Re-downloaded the bioRxiv PDF: its hash differs because of the per-request banner, but the pdftotext -layout text layer hashes exactly to the recorded value, so the content checked is the content extracted. Parsed all 279 cells (Table 2: 108, Table 3: 108, Table 4: 63) with an independent parser, mapping every result through its evaluation, configuration and protocol to a model, split and metric cell: all 279 printed and numeric values and all locator line numbers match, with no cell missed. Read 44 sampled values covering all three tables and all nine models from rendered pages 6, 10 and 12: 44 of 44 match. claims.csv rows, printed values and locators equal the batch.
numeric value: 0.392
metric qualifier: PLABench chembl35
missing metadata: uncertainty: reason: unreported

Related records

evaluation: FLOWR.ROOT on PLABench chembl35

Suggest a correction