Source

The role of protein embeddings for protein-protein interaction prediction with graph neural networks

Evidence concern: excluded from comparisons

The text's summary does not match the tables: it lists ProtT5 among structure embeddings while the tables print ProstT5; it reports a GraphSAGE with DistMult result (+38% Hits@100 on HuRI) although DistMult is not a column; and the stated gains (for example ESM-2 with GraphSAGE +6% Hits@100 on OGBl-ppa) cannot be reproduced from the printed table values (GS random 0.0104, ESM-2 0.2421). Values are kept as printed; comparisons should rely on the tables only after a reviewer has judged this. Random-feature baselines on OGBl-ppa (GCN 0.0588, GS 0.0104 Hits@100) look far below the public OGB leaderboard values the authors say they reproduced (Supplementary Table S2, not retrieved; reviewer did not verify the leaderboard values).The role of protein embeddings for protein-protein interaction prediction with graph neural networks · Results and discussion, first paragraph; Tables 5 and 6

Briefings in Bioinformatics 2026, bbag313; PMC13280947 full-text XML · Peer-reviewed

DOI: 10.1093/bib/bbag313

What this source contributed

Records in this release that cite this source. A record can cite several sources.

3 models
11 methods
77 configurations
1 benchmark
4 protocols
2 datasets
308 evaluations
308 results

Results

Values as the source printed them. Compare results only under compatible protocols, datasets and metrics.

Value	Metric	Model or configuration	Dataset	Evaluation
0.7917 fraction	Accuracy (HuRI PPI link prediction)	GCN with ComplEx protein embeddings (function (knowledge-graph embedding)) (PPI embeddings study)	HuRI PPI graph (embedding study)	GCN / ComplEx on HuRI (Accuracy)
0.6064 fraction	Hits at 100 (HuRI PPI link prediction)	GCN with ComplEx protein embeddings (function (knowledge-graph embedding)) (PPI embeddings study)	HuRI PPI graph (embedding study)	GCN / ComplEx on HuRI (Hits@100)
0.9842 fraction	Accuracy (OGBl-ppa PPI link prediction)	GCN with ComplEx protein embeddings (function (knowledge-graph embedding)) (PPI embeddings study)	OGBl-ppa PPI graph (embedding study)	GCN / ComplEx on OGBl-ppa (Accuracy)
0.2449 fraction	Hits at 100 (OGBl-ppa PPI link prediction)	GCN with ComplEx protein embeddings (function (knowledge-graph embedding)) (PPI embeddings study)	OGBl-ppa PPI graph (embedding study)	GCN / ComplEx on OGBl-ppa (Hits@100)
0.6715 fraction	Accuracy (HuRI PPI link prediction)	GCN with majority vote across the representations (PPI embeddings study)	HuRI PPI graph (embedding study)	GCN / Ensemble on HuRI (Accuracy)
0.5286 fraction	Hits at 100 (HuRI PPI link prediction)	GCN with majority vote across the representations (PPI embeddings study)	HuRI PPI graph (embedding study)	GCN / Ensemble on HuRI (Hits@100)
0.9416 fraction	Accuracy (OGBl-ppa PPI link prediction)	GCN with majority vote across the representations (PPI embeddings study)	OGBl-ppa PPI graph (embedding study)	GCN / Ensemble on OGBl-ppa (Accuracy)
0.2101 fraction	Hits at 100 (OGBl-ppa PPI link prediction)	GCN with majority vote across the representations (PPI embeddings study)	OGBl-ppa PPI graph (embedding study)	GCN / Ensemble on OGBl-ppa (Hits@100)
0.7018 fraction	Accuracy (HuRI PPI link prediction)	GCN with ESM-2 protein embeddings (sequence) (PPI embeddings study)	HuRI PPI graph (embedding study)	GCN / ESM-2 on HuRI (Accuracy)
0.4438 fraction	Hits at 100 (HuRI PPI link prediction)	GCN with ESM-2 protein embeddings (sequence) (PPI embeddings study)	HuRI PPI graph (embedding study)	GCN / ESM-2 on HuRI (Hits@100)
0.8729 fraction	Accuracy (OGBl-ppa PPI link prediction)	GCN with ESM-2 protein embeddings (sequence) (PPI embeddings study)	OGBl-ppa PPI graph (embedding study)	GCN / ESM-2 on OGBl-ppa (Accuracy)
0.1604 fraction	Hits at 100 (OGBl-ppa PPI link prediction)	GCN with ESM-2 protein embeddings (sequence) (PPI embeddings study)	OGBl-ppa PPI graph (embedding study)	GCN / ESM-2 on OGBl-ppa (Hits@100)
0.7602 fraction	Accuracy (HuRI PPI link prediction)	GCN with ESM-IF1 protein embeddings (structure) (PPI embeddings study)	HuRI PPI graph (embedding study)	GCN / ESM-IF1 on HuRI (Accuracy)
0.4968 fraction	Hits at 100 (HuRI PPI link prediction)	GCN with ESM-IF1 protein embeddings (structure) (PPI embeddings study)	HuRI PPI graph (embedding study)	GCN / ESM-IF1 on HuRI (Hits@100)
0.9814 fraction	Accuracy (OGBl-ppa PPI link prediction)	GCN with ESM-IF1 protein embeddings (structure) (PPI embeddings study)	OGBl-ppa PPI graph (embedding study)	GCN / ESM-IF1 on OGBl-ppa (Accuracy)
0.1851 fraction	Hits at 100 (OGBl-ppa PPI link prediction)	GCN with ESM-IF1 protein embeddings (structure) (PPI embeddings study)	OGBl-ppa PPI graph (embedding study)	GCN / ESM-IF1 on OGBl-ppa (Hits@100)
0.7044 fraction	Accuracy (HuRI PPI link prediction)	GCN with PiFold protein embeddings (structure) (PPI embeddings study)	HuRI PPI graph (embedding study)	GCN / PiFold on HuRI (Accuracy)
0.4573 fraction	Hits at 100 (HuRI PPI link prediction)	GCN with PiFold protein embeddings (structure) (PPI embeddings study)	HuRI PPI graph (embedding study)	GCN / PiFold on HuRI (Hits@100)
0.8988 fraction	Accuracy (OGBl-ppa PPI link prediction)	GCN with PiFold protein embeddings (structure) (PPI embeddings study)	OGBl-ppa PPI graph (embedding study)	GCN / PiFold on OGBl-ppa (Accuracy)
0.1893 fraction	Hits at 100 (OGBl-ppa PPI link prediction)	GCN with PiFold protein embeddings (structure) (PPI embeddings study)	OGBl-ppa PPI graph (embedding study)	GCN / PiFold on OGBl-ppa (Hits@100)
0.7154 fraction	Accuracy (HuRI PPI link prediction)	GCN with ProGen2 protein embeddings (sequence) (PPI embeddings study)	HuRI PPI graph (embedding study)	GCN / ProGen2 on HuRI (Accuracy)
0.4538 fraction	Hits at 100 (HuRI PPI link prediction)	GCN with ProGen2 protein embeddings (sequence) (PPI embeddings study)	HuRI PPI graph (embedding study)	GCN / ProGen2 on HuRI (Hits@100)
0.9128 fraction	Accuracy (OGBl-ppa PPI link prediction)	GCN with ProGen2 protein embeddings (sequence) (PPI embeddings study)	OGBl-ppa PPI graph (embedding study)	GCN / ProGen2 on OGBl-ppa (Accuracy)
0.2178 fraction	Hits at 100 (OGBl-ppa PPI link prediction)	GCN with ProGen2 protein embeddings (sequence) (PPI embeddings study)	OGBl-ppa PPI graph (embedding study)	GCN / ProGen2 on OGBl-ppa (Hits@100)
0.6497 fraction	Accuracy (HuRI PPI link prediction)	GCN with ProstT5 protein embeddings (structure) (PPI embeddings study)	HuRI PPI graph (embedding study)	GCN / ProstT5 on HuRI (Accuracy)
0.4470 fraction	Hits at 100 (HuRI PPI link prediction)	GCN with ProstT5 protein embeddings (structure) (PPI embeddings study)	HuRI PPI graph (embedding study)	GCN / ProstT5 on HuRI (Hits@100)
0.8081 fraction	Accuracy (OGBl-ppa PPI link prediction)	GCN with ProstT5 protein embeddings (structure) (PPI embeddings study)	OGBl-ppa PPI graph (embedding study)	GCN / ProstT5 on OGBl-ppa (Accuracy)
0.1616 fraction	Hits at 100 (OGBl-ppa PPI link prediction)	GCN with ProstT5 protein embeddings (structure) (PPI embeddings study)	OGBl-ppa PPI graph (embedding study)	GCN / ProstT5 on OGBl-ppa (Hits@100)
0.6750 fraction	Accuracy (HuRI PPI link prediction)	GCN with ProtTrans protein embeddings (sequence) (PPI embeddings study)	HuRI PPI graph (embedding study)	GCN / ProtTrans on HuRI (Accuracy)
0.4501 fraction	Hits at 100 (HuRI PPI link prediction)	GCN with ProtTrans protein embeddings (sequence) (PPI embeddings study)	HuRI PPI graph (embedding study)	GCN / ProtTrans on HuRI (Hits@100)
0.8714 fraction	Accuracy (OGBl-ppa PPI link prediction)	GCN with ProtTrans protein embeddings (sequence) (PPI embeddings study)	OGBl-ppa PPI graph (embedding study)	GCN / ProtTrans on OGBl-ppa (Accuracy)
0.1677 fraction	Hits at 100 (OGBl-ppa PPI link prediction)	GCN with ProtTrans protein embeddings (sequence) (PPI embeddings study)	OGBl-ppa PPI graph (embedding study)	GCN / ProtTrans on OGBl-ppa (Hits@100)
0.6995 fraction	Accuracy (HuRI PPI link prediction)	GCN with randomly initialised node features (baseline) (PPI embeddings study)	HuRI PPI graph (embedding study)	GCN / Random on HuRI (Accuracy)
0.4350 fraction	Hits at 100 (HuRI PPI link prediction)	GCN with randomly initialised node features (baseline) (PPI embeddings study)	HuRI PPI graph (embedding study)	GCN / Random on HuRI (Hits@100)
0.8672 fraction	Accuracy (OGBl-ppa PPI link prediction)	GCN with randomly initialised node features (baseline) (PPI embeddings study)	OGBl-ppa PPI graph (embedding study)	GCN / Random on OGBl-ppa (Accuracy)
0.0588 fraction	Hits at 100 (OGBl-ppa PPI link prediction)	GCN with randomly initialised node features (baseline) (PPI embeddings study)	OGBl-ppa PPI graph (embedding study)	GCN / Random on OGBl-ppa (Hits@100)
0.7989 fraction	Accuracy (HuRI PPI link prediction)	GCN with RDF2Vec protein embeddings (function (knowledge-graph embedding)) (PPI embeddings study)	HuRI PPI graph (embedding study)	GCN / RDF2Vec on HuRI (Accuracy)
0.6508 fraction	Hits at 100 (HuRI PPI link prediction)	GCN with RDF2Vec protein embeddings (function (knowledge-graph embedding)) (PPI embeddings study)	HuRI PPI graph (embedding study)	GCN / RDF2Vec on HuRI (Hits@100)
0.9816 fraction	Accuracy (OGBl-ppa PPI link prediction)	GCN with RDF2Vec protein embeddings (function (knowledge-graph embedding)) (PPI embeddings study)	OGBl-ppa PPI graph (embedding study)	GCN / RDF2Vec on OGBl-ppa (Accuracy)
0.2553 fraction	Hits at 100 (OGBl-ppa PPI link prediction)	GCN with RDF2Vec protein embeddings (function (knowledge-graph embedding)) (PPI embeddings study)	OGBl-ppa PPI graph (embedding study)	GCN / RDF2Vec on OGBl-ppa (Hits@100)
0.5150 fraction	Accuracy (HuRI PPI link prediction)	GCN with TransE protein embeddings (function (knowledge-graph embedding)) (PPI embeddings study)	HuRI PPI graph (embedding study)	GCN / TransE on HuRI (Accuracy)
0.4318 fraction	Hits at 100 (HuRI PPI link prediction)	GCN with TransE protein embeddings (function (knowledge-graph embedding)) (PPI embeddings study)	HuRI PPI graph (embedding study)	GCN / TransE on HuRI (Hits@100)
0.9010 fraction	Accuracy (OGBl-ppa PPI link prediction)	GCN with TransE protein embeddings (function (knowledge-graph embedding)) (PPI embeddings study)	OGBl-ppa PPI graph (embedding study)	GCN / TransE on OGBl-ppa (Accuracy)
0.0382 fraction	Hits at 100 (OGBl-ppa PPI link prediction)	GCN with TransE protein embeddings (function (knowledge-graph embedding)) (PPI embeddings study)	OGBl-ppa PPI graph (embedding study)	GCN / TransE on OGBl-ppa (Hits@100)
0.7717 fraction	Accuracy (HuRI PPI link prediction)	GraphSAGE with ComplEx protein embeddings (function (knowledge-graph embedding)) (PPI embeddings study)	HuRI PPI graph (embedding study)	GraphSAGE / ComplEx on HuRI (Accuracy)
0.5622 fraction	Hits at 100 (HuRI PPI link prediction)	GraphSAGE with ComplEx protein embeddings (function (knowledge-graph embedding)) (PPI embeddings study)	HuRI PPI graph (embedding study)	GraphSAGE / ComplEx on HuRI (Hits@100)
0.9741 fraction	Accuracy (OGBl-ppa PPI link prediction)	GraphSAGE with ComplEx protein embeddings (function (knowledge-graph embedding)) (PPI embeddings study)	OGBl-ppa PPI graph (embedding study)	GraphSAGE / ComplEx on OGBl-ppa (Accuracy)
0.2758 fraction	Hits at 100 (OGBl-ppa PPI link prediction)	GraphSAGE with ComplEx protein embeddings (function (knowledge-graph embedding)) (PPI embeddings study)	OGBl-ppa PPI graph (embedding study)	GraphSAGE / ComplEx on OGBl-ppa (Hits@100)
0.6221 fraction	Accuracy (HuRI PPI link prediction)	GraphSAGE with majority vote across the representations (PPI embeddings study)	HuRI PPI graph (embedding study)	GraphSAGE / Ensemble on HuRI (Accuracy)
0.3121 fraction	Hits at 100 (HuRI PPI link prediction)	GraphSAGE with majority vote across the representations (PPI embeddings study)	HuRI PPI graph (embedding study)	GraphSAGE / Ensemble on HuRI (Hits@100)

Models, benchmarks and other records

Models (3)

Methods (11)

Configurations (77)

Benchmarks (1)

Protein embeddings for GNN PPI prediction (OGBl-ppa and HuRI)

Protocols (4)

Datasets (2)

Evaluations (308)

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.

24 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
attributes.artifact_sha256 360d4f17b779e07f170c3f0c3755bfbfddd064831bf5a81d2642a4cfffd4866c Source metadata	The role of protein embeddings for protein-protein interaction prediction with graph neural networks Original source ↗ No field-specific location recorded Version: Briefings in Bioinformatics 2026, bbag313; PMC13280947 full-text XML Retrieved: 2026-10-11T21:27:23Z	catalogued No individual claim review recorded Source has a recorded evidence concern. Consult its source page before using the claim. Audit details Field: `attributes.artifact_sha256` Source artifact SHA-256: `360d4f17b779e07f170c3f0c3755bfbfddd064831bf5a81d2642a4cfffd4866c` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
Artifact url https://www.ebi.ac.uk/europepmc/webservices/rest/PMC13280947/fullTextXML Source metadata	The role of protein embeddings for protein-protein interaction prediction with graph neural networks Original source ↗ No field-specific location recorded Version: Briefings in Bioinformatics 2026, bbag313; PMC13280947 full-text XML Retrieved: 2026-10-11T21:27:23Z	catalogued No individual claim review recorded Source has a recorded evidence concern. Consult its source page before using the claim. Audit details Field: `attributes.artifact_url` Source artifact SHA-256: `360d4f17b779e07f170c3f0c3755bfbfddd064831bf5a81d2642a4cfffd4866c` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
Doi 10.1093/bib/bbag313 Source metadata	The role of protein embeddings for protein-protein interaction prediction with graph neural networks Original source ↗ No field-specific location recorded Version: Briefings in Bioinformatics 2026, bbag313; PMC13280947 full-text XML Retrieved: 2026-10-11T21:27:23Z	catalogued No individual claim review recorded Source has a recorded evidence concern. Consult its source page before using the claim. Audit details Field: `attributes.doi` Source artifact SHA-256: `360d4f17b779e07f170c3f0c3755bfbfddd064831bf5a81d2642a4cfffd4866c` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
Evidence concerns 1 values source id lit-interactions-20261011-ppiemb-source-bbag313 artifact sha256 360d4f17b779e07f170c3f0c3755bfbfddd064831bf5a81d2642a4cfffd4866c source locator Results and discussion, first paragraph; Tables 5 and 6 message The text's summary does not match the tables: it lists ProtT5 among structure embeddings while the tables print ProstT5; it reports a GraphSAGE with DistMult result (+38% Hits@100 on HuRI) although DistMult is not a column; and the stated gains (for example ESM-2 with GraphSAGE +6% Hits@100 on OGBl-ppa) cannot be reproduced from the printed table values (GS random 0.0104, ESM-2 0.2421). Values are kept as printed; comparisons should rely on the tables only after a reviewer has judged this. Random-feature baselines on OGBl-ppa (GCN 0.0588, GS 0.0104 Hits@100) look far below the public OGB leaderboard values the authors say they reproduced (Supplementary Table S2, not retrieved; reviewer did not verify the leaderboard values). reviewed at 2026-10-11T21:40:00Z review method ai-assisted-source-review Source metadata	The role of protein embeddings for protein-protein interaction prediction with graph neural networks Original source ↗ No field-specific location recorded Version: Briefings in Bioinformatics 2026, bbag313; PMC13280947 full-text XML Retrieved: 2026-10-11T21:27:23Z	catalogued No individual claim review recorded Source has a recorded evidence concern. Consult its source page before using the claim. Audit details Field: `attributes.evidence_concerns` Source artifact SHA-256: `360d4f17b779e07f170c3f0c3755bfbfddd064831bf5a81d2642a4cfffd4866c` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
Licence CC-BY-NC-4.0 Source metadata	The role of protein embeddings for protein-protein interaction prediction with graph neural networks Original source ↗ No field-specific location recorded Version: Briefings in Bioinformatics 2026, bbag313; PMC13280947 full-text XML Retrieved: 2026-10-11T21:27:23Z	catalogued No individual claim review recorded Source has a recorded evidence concern. Consult its source page before using the claim. Audit details Field: `attributes.licence` Source artifact SHA-256: `360d4f17b779e07f170c3f0c3755bfbfddd064831bf5a81d2642a4cfffd4866c` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
Limitations 2 values The supplementary material (Tables S1-S7: datasets, original OGBl-ppa leaderboard values, embedding and GNN parameters) was not retrieved. No per-cell uncertainty or run count is printed; the footnote of Tables 5 and 6 gives only the table-wide maximum standard deviation (0.0711 OGBl-ppa, 0.0928 HuRI). Source metadata	The role of protein embeddings for protein-protein interaction prediction with graph neural networks Original source ↗ No field-specific location recorded Version: Briefings in Bioinformatics 2026, bbag313; PMC13280947 full-text XML Retrieved: 2026-10-11T21:27:23Z	catalogued No individual claim review recorded Source has a recorded evidence concern. Consult its source page before using the claim. Audit details Field: `attributes.limitations` Source artifact SHA-256: `360d4f17b779e07f170c3f0c3755bfbfddd064831bf5a81d2642a4cfffd4866c` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
Media type application/xml Source metadata	The role of protein embeddings for protein-protein interaction prediction with graph neural networks Original source ↗ No field-specific location recorded Version: Briefings in Bioinformatics 2026, bbag313; PMC13280947 full-text XML Retrieved: 2026-10-11T21:27:23Z	catalogued No individual claim review recorded Source has a recorded evidence concern. Consult its source page before using the claim. Audit details Field: `attributes.media_type` Source artifact SHA-256: `360d4f17b779e07f170c3f0c3755bfbfddd064831bf5a81d2642a4cfffd4866c` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
Publication status peer_reviewed Source metadata	The role of protein embeddings for protein-protein interaction prediction with graph neural networks Original source ↗ No field-specific location recorded Version: Briefings in Bioinformatics 2026, bbag313; PMC13280947 full-text XML Retrieved: 2026-10-11T21:27:23Z	catalogued No individual claim review recorded Source has a recorded evidence concern. Consult its source page before using the claim. Audit details Field: `attributes.publication_status` Source artifact SHA-256: `360d4f17b779e07f170c3f0c3755bfbfddd064831bf5a81d2642a4cfffd4866c` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
Retrieved at 2026-10-11T21:27:23Z Source metadata	The role of protein embeddings for protein-protein interaction prediction with graph neural networks Original source ↗ No field-specific location recorded Version: Briefings in Bioinformatics 2026, bbag313; PMC13280947 full-text XML Retrieved: 2026-10-11T21:27:23Z	catalogued No individual claim review recorded Source has a recorded evidence concern. Consult its source page before using the claim. Audit details Field: `attributes.retrieved_at` Source artifact SHA-256: `360d4f17b779e07f170c3f0c3755bfbfddd064831bf5a81d2642a4cfffd4866c` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
Review: date 2026-10-11 Source metadata	The role of protein embeddings for protein-protein interaction prediction with graph neural networks Original source ↗ No field-specific location recorded Version: Briefings in Bioinformatics 2026, bbag313; PMC13280947 full-text XML Retrieved: 2026-10-11T21:27:23Z	catalogued No individual claim review recorded Source has a recorded evidence concern. Consult its source page before using the claim. Audit details Field: `attributes.review.date` Source artifact SHA-256: `360d4f17b779e07f170c3f0c3755bfbfddd064831bf5a81d2642a4cfffd4866c` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact

How to read the evidence

Sources and history

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

Primary source retrieved and hashed for the 2026-10-11 microbial and interactions literature pass.

Release history

No supporting source is linked yet.

Read original source

Download this release (gzip)

Technical metadata and extraction receipts

Stable ID: lit-interactions-20261011-ppiemb-source-bbag313

areas: molecular-interactions; proteins-complexes
contexts: research
url: https://doi.org/10.1093/bib/bbag313
artifact url: https://www.ebi.ac.uk/europepmc/webservices/rest/PMC13280947/fullTextXML
version: Briefings in Bioinformatics 2026, bbag313; PMC13280947 full-text XML
retrieved at: 2026-10-11T21:27:23Z
artifact sha256: 360d4f17b779e07f170c3f0c3755bfbfddd064831bf5a81d2642a4cfffd4866c
doi: 10.1093/bib/bbag313
publication status: peer_reviewed
licence: CC-BY-NC-4.0
media type: application/xml
limitations: The supplementary material (Tables S1-S7: datasets, original OGBl-ppa leaderboard values, embedding and GNN parameters) was not retrieved.; No per-cell uncertainty or run count is printed; the footnote of Tables 5 and 6 gives only the table-wide maximum standard deviation (0.0711 OGBl-ppa, 0.0928 HuRI).
evidence concerns: source id: lit-interactions-20261011-ppiemb-source-bbag313; artifact sha256: 360d4f17b779e07f170c3f0c3755bfbfddd064831bf5a81d2642a4cfffd4866c; source locator: Results and discussion, first paragraph; Tables 5 and 6; message: The text's summary does not match the tables: it lists ProtT5 among structure embeddings while the tables print ProstT5; it reports a GraphSAGE with DistMult result (+38% Hits@100 on HuRI) although DistMult is not a column; and the stated gains (for example ESM-2 with GraphSAGE +6% Hits@100 on OGBl-ppa) cannot be reproduced from the printed table values (GS random 0.0104, ESM-2 0.2421). Values are kept as printed; comparisons should rely on the tables only after a reviewer has judged this. Random-feature baselines on OGBl-ppa (GCN 0.0588, GS 0.0104 Hits@100) look far below the public OGB leaderboard values the authors say they reproduced (Supplementary Table S2, not retrieved; reviewer did not verify the leaderboard values).; reviewed at: 2026-10-11T21:40:00Z; review method: ai-assisted-source-review
review: method: source-hash-verification; method note: Full-text XML re-downloaded from the recorded Europe PMC URL and its SHA-256 matched.; 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

Related records

Suggest a correction