rewirebio.iobenchmarks
← Browse all sources
Source

Genome-wide prediction of disease variant effects with a deep protein language model (Brandes et al., Nature Genetics 2023)

Nature Genetics · 2023 · Version of record, Nature Genetics 55, 1512-1522 (published 2023-08-10) · Peer-reviewed

DOI: 10.1038/s41588-023-01465-0

What this source contributed

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

Results

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

ValueMetricModel or configurationDatasetEvaluation
0.662203060682712 unitlessAUROC (ROC-AUC, pathogenic versus benign ClinVar all stop-gains; zero-shot)50 bp rule for nonsense-mediated decayClinVar stop-gain variants, pathogenic vs benign (Brandes et al. 2023)50bp rule on all stop-gains (Brandes et al. 2023)
0.586435752317274 unitlessAUROC (ROC-AUC, pathogenic versus benign ClinVar all in-frame indels; zero-shot)Pairwise alignment score (indel baseline)ClinVar in-frame indels, pathogenic vs benign (Brandes et al. 2023)Alignment on all in-frame indels (Brandes et al. 2023)
0.946487950167752 unitlessAUPRC (balanced PRC-AUC over the variants both methods score: mean over 100 random downsamples to equal pathogenic and benign counts (80% of the minority class); zero-shot)BayesDel (no AF) (Brandes et al. 2023 head-to-head comparison)ClinVar missense variants scored by both ESM1b and BayesDel (no AF) (Brandes et al. 2023)BayesDel (no AF) on ClinVar missense, head to head against ESM1b (Brandes et al. 2023)
0.947999495980187 unitlessAUROC (global ROC-AUC over the variants both methods score; zero-shot)BayesDel (no AF) (Brandes et al. 2023 head-to-head comparison)ClinVar missense variants scored by both ESM1b and BayesDel (no AF) (Brandes et al. 2023)BayesDel (no AF) on ClinVar missense, head to head against ESM1b (Brandes et al. 2023)
0.980335926822551 unitlessAUPRC (balanced PRC-AUC over the variants both methods score: mean over 100 random downsamples to equal pathogenic and benign counts (80% of the minority class); zero-shot)BayesDel (no AF) (Brandes et al. 2023 head-to-head comparison)HGMD/gnomAD missense variants scored by both ESM1b and BayesDel (no AF) (Brandes et al. 2023)BayesDel (no AF) on HGMD/gnomAD missense, head to head against ESM1b (Brandes et al. 2023)
0.978236677449422 unitlessAUROC (global ROC-AUC over the variants both methods score; zero-shot)BayesDel (no AF) (Brandes et al. 2023 head-to-head comparison)HGMD/gnomAD missense variants scored by both ESM1b and BayesDel (no AF) (Brandes et al. 2023)BayesDel (no AF) on HGMD/gnomAD missense, head to head against ESM1b (Brandes et al. 2023)
0.974160079812455 unitlessAUPRC (balanced PRC-AUC over the variants both methods score: mean over 100 random downsamples to equal pathogenic and benign counts (80% of the minority class); zero-shot)BayesDel (with AF) (Brandes et al. 2023 head-to-head comparison)ClinVar missense variants scored by both ESM1b and BayesDel (with AF) (Brandes et al. 2023)BayesDel (with AF) on ClinVar missense, head to head against ESM1b (Brandes et al. 2023)
0.976738413063963 unitlessAUROC (global ROC-AUC over the variants both methods score; zero-shot)BayesDel (with AF) (Brandes et al. 2023 head-to-head comparison)ClinVar missense variants scored by both ESM1b and BayesDel (with AF) (Brandes et al. 2023)BayesDel (with AF) on ClinVar missense, head to head against ESM1b (Brandes et al. 2023)
0.992324635108394 unitlessAUPRC (balanced PRC-AUC over the variants both methods score: mean over 100 random downsamples to equal pathogenic and benign counts (80% of the minority class); zero-shot)BayesDel (with AF) (Brandes et al. 2023 head-to-head comparison)HGMD/gnomAD missense variants scored by both ESM1b and BayesDel (with AF) (Brandes et al. 2023)BayesDel (with AF) on HGMD/gnomAD missense, head to head against ESM1b (Brandes et al. 2023)
0.992291705912356 unitlessAUROC (global ROC-AUC over the variants both methods score; zero-shot)BayesDel (with AF) (Brandes et al. 2023 head-to-head comparison)HGMD/gnomAD missense variants scored by both ESM1b and BayesDel (with AF) (Brandes et al. 2023)BayesDel (with AF) on HGMD/gnomAD missense, head to head against ESM1b (Brandes et al. 2023)
0.581684812122288 unitlessAUROC (ROC-AUC, pathogenic versus benign ClinVar all in-frame indels; zero-shot)BlastP score (indel baseline)ClinVar in-frame indels, pathogenic vs benign (Brandes et al. 2023)BlastP on all in-frame indels (Brandes et al. 2023)
0.539298346796558 unitlessAUPRC (balanced PRC-AUC over the variants both methods score: mean over 100 random downsamples to equal pathogenic and benign counts (80% of the minority class); zero-shot)bStatistic (Brandes et al. 2023 clinical benchmarks)ClinVar missense variants scored by both ESM1b and bStatistic (Brandes et al. 2023)bStatistic on ClinVar missense, head to head against ESM1b (Brandes et al. 2023)
0.560656139972137 unitlessAUROC (global ROC-AUC over the variants both methods score; zero-shot)bStatistic (Brandes et al. 2023 clinical benchmarks)ClinVar missense variants scored by both ESM1b and bStatistic (Brandes et al. 2023)bStatistic on ClinVar missense, head to head against ESM1b (Brandes et al. 2023)
0.543147322509533 unitlessAUPRC (balanced PRC-AUC: mean over 100 random downsamples to equal pathogenic and benign counts (80% of the minority class); zero-shot)bStatistic (Brandes et al. 2023 clinical benchmarks)ClinVar missense pathogenic vs benign, primary isoforms (Brandes et al. 2023, from the EVE portal)bStatistic on ClinVar missense, variants scored by all 19 methods (Brandes et al. 2023)
0.562825728589234 unitlessAUROC (global ROC-AUC over all variants; zero-shot)bStatistic (Brandes et al. 2023 clinical benchmarks)ClinVar missense pathogenic vs benign, primary isoforms (Brandes et al. 2023, from the EVE portal)bStatistic on ClinVar missense, variants scored by all 19 methods (Brandes et al. 2023)
0.580555088476076 unitlessAUPRC (balanced PRC-AUC over the variants both methods score: mean over 100 random downsamples to equal pathogenic and benign counts (80% of the minority class); zero-shot)bStatistic (Brandes et al. 2023 clinical benchmarks)HGMD/gnomAD missense variants scored by both ESM1b and bStatistic (Brandes et al. 2023)bStatistic on HGMD/gnomAD missense, head to head against ESM1b (Brandes et al. 2023)
0.581782069823495 unitlessAUROC (global ROC-AUC over the variants both methods score; zero-shot)bStatistic (Brandes et al. 2023 clinical benchmarks)HGMD/gnomAD missense variants scored by both ESM1b and bStatistic (Brandes et al. 2023)bStatistic on HGMD/gnomAD missense, head to head against ESM1b (Brandes et al. 2023)
0.534115963003298 unitlessAUPRC (balanced PRC-AUC: mean over 100 random downsamples to equal pathogenic and benign counts (80% of the minority class); zero-shot)bStatistic (Brandes et al. 2023 clinical benchmarks)HGMD disease-causing vs gnomAD common missense variants, primary isoforms (Brandes et al. 2023)bStatistic on HGMD/gnomAD missense, variants scored by all 19 methods (Brandes et al. 2023)
0.535573966000815 unitlessAUROC (global ROC-AUC over all variants; zero-shot)bStatistic (Brandes et al. 2023 clinical benchmarks)HGMD disease-causing vs gnomAD common missense variants, primary isoforms (Brandes et al. 2023)bStatistic on HGMD/gnomAD missense, variants scored by all 19 methods (Brandes et al. 2023)
0.835780384285267 unitlessAUROC (ROC-AUC, pathogenic versus benign ClinVar all in-frame indels; zero-shot)CADD in-frame indel scores (Brandes et al. 2023)ClinVar in-frame indels, pathogenic vs benign (Brandes et al. 2023)CADD on all in-frame indels (Brandes et al. 2023)
0.871368321583246 unitlessAUPRC (balanced PRC-AUC over the variants both methods score: mean over 100 random downsamples to equal pathogenic and benign counts (80% of the minority class); zero-shot)CADD (Brandes et al. 2023 head-to-head comparison)ClinVar missense variants scored by both ESM1b and CADD (Brandes et al. 2023)CADD on ClinVar missense, head to head against ESM1b (Brandes et al. 2023)
0.896607058523481 unitlessAUROC (global ROC-AUC over the variants both methods score; zero-shot)CADD (Brandes et al. 2023 head-to-head comparison)ClinVar missense variants scored by both ESM1b and CADD (Brandes et al. 2023)CADD on ClinVar missense, head to head against ESM1b (Brandes et al. 2023)
0.671464 unitlessAUROC (ROC-AUC, pathogenic versus benign ClinVar deletion-insertions; zero-shot)CADD in-frame indel scores (Brandes et al. 2023)ClinVar in-frame deletion-insertions (Brandes et al. 2023)CADD on deletion-insertions (Brandes et al. 2023)
0.848141 unitlessAUROC (ROC-AUC, pathogenic versus benign ClinVar deletions; zero-shot)CADD in-frame indel scores (Brandes et al. 2023)ClinVar in-frame deletions (Brandes et al. 2023)CADD on deletions (Brandes et al. 2023)
0.911693982661796 unitlessAUPRC (balanced PRC-AUC over the variants both methods score: mean over 100 random downsamples to equal pathogenic and benign counts (80% of the minority class); zero-shot)CADD (Brandes et al. 2023 head-to-head comparison)HGMD/gnomAD missense variants scored by both ESM1b and CADD (Brandes et al. 2023)CADD on HGMD/gnomAD missense, head to head against ESM1b (Brandes et al. 2023)
0.917203727649111 unitlessAUROC (global ROC-AUC over the variants both methods score; zero-shot)CADD (Brandes et al. 2023 head-to-head comparison)HGMD/gnomAD missense variants scored by both ESM1b and CADD (Brandes et al. 2023)CADD on HGMD/gnomAD missense, head to head against ESM1b (Brandes et al. 2023)
0.756837 unitlessAUROC (ROC-AUC, pathogenic versus benign ClinVar insertions; zero-shot)CADD in-frame indel scores (Brandes et al. 2023)ClinVar in-frame insertions (Brandes et al. 2023)CADD on insertions (Brandes et al. 2023)
0.978923571976287 unitlessAUPRC (balanced PRC-AUC over the variants both methods score: mean over 100 random downsamples to equal pathogenic and benign counts (80% of the minority class); zero-shot)ClinPred (Brandes et al. 2023 head-to-head comparison)ClinVar missense variants scored by both ESM1b and ClinPred (Brandes et al. 2023)ClinPred on ClinVar missense, head to head against ESM1b (Brandes et al. 2023)
0.98197198740626 unitlessAUROC (global ROC-AUC over the variants both methods score; zero-shot)ClinPred (Brandes et al. 2023 head-to-head comparison)ClinVar missense variants scored by both ESM1b and ClinPred (Brandes et al. 2023)ClinPred on ClinVar missense, head to head against ESM1b (Brandes et al. 2023)
0.980436885355161 unitlessAUPRC (balanced PRC-AUC over the variants both methods score: mean over 100 random downsamples to equal pathogenic and benign counts (80% of the minority class); zero-shot)ClinPred (Brandes et al. 2023 head-to-head comparison)HGMD/gnomAD missense variants scored by both ESM1b and ClinPred (Brandes et al. 2023)ClinPred on HGMD/gnomAD missense, head to head against ESM1b (Brandes et al. 2023)
0.982613471236555 unitlessAUROC (global ROC-AUC over the variants both methods score; zero-shot)ClinPred (Brandes et al. 2023 head-to-head comparison)HGMD/gnomAD missense variants scored by both ESM1b and ClinPred (Brandes et al. 2023)ClinPred on HGMD/gnomAD missense, head to head against ESM1b (Brandes et al. 2023)
0.723404287845123 unitlessAUPRC (balanced PRC-AUC over the variants both methods score: mean over 100 random downsamples to equal pathogenic and benign counts (80% of the minority class); zero-shot)DANN (Brandes et al. 2023 head-to-head comparison)ClinVar missense variants scored by both ESM1b and DANN (Brandes et al. 2023)DANN on ClinVar missense, head to head against ESM1b (Brandes et al. 2023)
0.779479521810876 unitlessAUROC (global ROC-AUC over the variants both methods score; zero-shot)DANN (Brandes et al. 2023 head-to-head comparison)ClinVar missense variants scored by both ESM1b and DANN (Brandes et al. 2023)DANN on ClinVar missense, head to head against ESM1b (Brandes et al. 2023)
0.792631936110773 unitlessAUPRC (balanced PRC-AUC over the variants both methods score: mean over 100 random downsamples to equal pathogenic and benign counts (80% of the minority class); zero-shot)DANN (Brandes et al. 2023 head-to-head comparison)HGMD/gnomAD missense variants scored by both ESM1b and DANN (Brandes et al. 2023)DANN on HGMD/gnomAD missense, head to head against ESM1b (Brandes et al. 2023)
0.838630621790844 unitlessAUROC (global ROC-AUC over the variants both methods score; zero-shot)DANN (Brandes et al. 2023 head-to-head comparison)HGMD/gnomAD missense variants scored by both ESM1b and DANN (Brandes et al. 2023)DANN on HGMD/gnomAD missense, head to head against ESM1b (Brandes et al. 2023)
0.917908571392221 unitlessAUPRC (balanced PRC-AUC over the variants both methods score: mean over 100 random downsamples to equal pathogenic and benign counts (80% of the minority class); zero-shot)DEOGEN2 (Brandes et al. 2023 head-to-head comparison)ClinVar missense variants scored by both ESM1b and DEOGEN2 (Brandes et al. 2023)DEOGEN2 on ClinVar missense, head to head against ESM1b (Brandes et al. 2023)
0.912478948636199 unitlessAUROC (global ROC-AUC over the variants both methods score; zero-shot)DEOGEN2 (Brandes et al. 2023 head-to-head comparison)ClinVar missense variants scored by both ESM1b and DEOGEN2 (Brandes et al. 2023)DEOGEN2 on ClinVar missense, head to head against ESM1b (Brandes et al. 2023)
0.961594714736039 unitlessAUPRC (balanced PRC-AUC over the variants both methods score: mean over 100 random downsamples to equal pathogenic and benign counts (80% of the minority class); zero-shot)DEOGEN2 (Brandes et al. 2023 head-to-head comparison)HGMD/gnomAD missense variants scored by both ESM1b and DEOGEN2 (Brandes et al. 2023)DEOGEN2 on HGMD/gnomAD missense, head to head against ESM1b (Brandes et al. 2023)
0.958144497005041 unitlessAUROC (global ROC-AUC over the variants both methods score; zero-shot)DEOGEN2 (Brandes et al. 2023 head-to-head comparison)HGMD/gnomAD missense variants scored by both ESM1b and DEOGEN2 (Brandes et al. 2023)DEOGEN2 on HGMD/gnomAD missense, head to head against ESM1b (Brandes et al. 2023)
0.501679198720091 unitlessAUROC (ROC-AUC, pathogenic versus benign ClinVar all in-frame indels; zero-shot)LevenshteinClinVar in-frame indels, pathogenic vs benign (Brandes et al. 2023)Edit distance on all in-frame indels (Brandes et al. 2023)
0.893881911468068 unitlessAUPRC (balanced PRC-AUC over the variants both methods score: mean over 100 random downsamples to equal pathogenic and benign counts (80% of the minority class); zero-shot)Eigen (Brandes et al. 2023 head-to-head comparison)ClinVar missense variants scored by both ESM1b and Eigen (Brandes et al. 2023)Eigen on ClinVar missense, head to head against ESM1b (Brandes et al. 2023)
0.902556806623812 unitlessAUROC (global ROC-AUC over the variants both methods score; zero-shot)Eigen (Brandes et al. 2023 head-to-head comparison)ClinVar missense variants scored by both ESM1b and Eigen (Brandes et al. 2023)Eigen on ClinVar missense, head to head against ESM1b (Brandes et al. 2023)
0.910438720140354 unitlessAUPRC (balanced PRC-AUC over the variants both methods score: mean over 100 random downsamples to equal pathogenic and benign counts (80% of the minority class); zero-shot)Eigen (Brandes et al. 2023 head-to-head comparison)HGMD/gnomAD missense variants scored by both ESM1b and Eigen (Brandes et al. 2023)Eigen on HGMD/gnomAD missense, head to head against ESM1b (Brandes et al. 2023)
0.913137960845037 unitlessAUROC (global ROC-AUC over the variants both methods score; zero-shot)Eigen (Brandes et al. 2023 head-to-head comparison)HGMD/gnomAD missense variants scored by both ESM1b and Eigen (Brandes et al. 2023)Eigen on HGMD/gnomAD missense, head to head against ESM1b (Brandes et al. 2023)
0.875445861039715 unitlessAUPRC (balanced PRC-AUC over the variants both methods score: mean over 100 random downsamples to equal pathogenic and benign counts (80% of the minority class); zero-shot)Eigen-PC (Brandes et al. 2023 head-to-head comparison)ClinVar missense variants scored by both ESM1b and Eigen-PC (Brandes et al. 2023)Eigen-PC on ClinVar missense, head to head against ESM1b (Brandes et al. 2023)
0.887878772403633 unitlessAUROC (global ROC-AUC over the variants both methods score; zero-shot)Eigen-PC (Brandes et al. 2023 head-to-head comparison)ClinVar missense variants scored by both ESM1b and Eigen-PC (Brandes et al. 2023)Eigen-PC on ClinVar missense, head to head against ESM1b (Brandes et al. 2023)
0.902808504951316 unitlessAUPRC (balanced PRC-AUC over the variants both methods score: mean over 100 random downsamples to equal pathogenic and benign counts (80% of the minority class); zero-shot)Eigen-PC (Brandes et al. 2023 head-to-head comparison)HGMD/gnomAD missense variants scored by both ESM1b and Eigen-PC (Brandes et al. 2023)Eigen-PC on HGMD/gnomAD missense, head to head against ESM1b (Brandes et al. 2023)
0.90771824730958 unitlessAUROC (global ROC-AUC over the variants both methods score; zero-shot)Eigen-PC (Brandes et al. 2023 head-to-head comparison)HGMD/gnomAD missense variants scored by both ESM1b and Eigen-PC (Brandes et al. 2023)Eigen-PC on HGMD/gnomAD missense, head to head against ESM1b (Brandes et al. 2023)
0.906188422497614 unitlessAUPRC (balanced PRC-AUC over the variants both methods score: mean over 100 random downsamples to equal pathogenic and benign counts (80% of the minority class); zero-shot)ESM1b (Brandes et al. 2023 clinical benchmarks)ClinVar missense variants scored by both ESM1b and BayesDel (no AF) (Brandes et al. 2023)ESM1b on ClinVar missense, head to head against BayesDel (no AF) (Brandes et al. 2023)
0.905875955210954 unitlessAUROC (global ROC-AUC over the variants both methods score; zero-shot)ESM1b (Brandes et al. 2023 clinical benchmarks)ClinVar missense variants scored by both ESM1b and BayesDel (no AF) (Brandes et al. 2023)ESM1b on ClinVar missense, head to head against BayesDel (no AF) (Brandes et al. 2023)

Models, benchmarks and other records

Methods (6)

Configurations (51)

Benchmarks (2)

Protocols (8)

Datasets (4)

Dataset subsets (5)

Evaluations (248)

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.

31 evidence rows matching the loaded filters

Claims, original sources and review scope · Release 2026-10-22-1b7cf027e4f6
Property and statementOriginal source and locationReview and provenance
Archive member
artifacts/brandes2023-pmc10484790.xml.gz
Source metadata		Genome-wide prediction of disease variant effects with a deep protein language model (Brandes et al., Nature Genetics 2023)

Original source ↗

No field-specific location recorded

Version: Version of record, Nature Genetics 55, 1512-1522 (published 2023-08-10)
Retrieved: 2026-10-14T12:03:14Z

catalogued

No individual claim review recorded

Audit details

Field: attributes.archive_member

Source artifact SHA-256: 919ca366369efbe7e0880565192de127375edddedb1aa43ad023dec612bb4a04

Hash scope: Hash scope not separately documented; inspect source record

Archive member: artifacts/brandes2023-pmc10484790.xml.gz

Inspected artifact

attributes.archive_sha256
0b3f30de21eff52626d21c249533e7e76e52685d1cfd02cc51b444668b1f58f0
Source metadata		Genome-wide prediction of disease variant effects with a deep protein language model (Brandes et al., Nature Genetics 2023)

Original source ↗

No field-specific location recorded

Version: Version of record, Nature Genetics 55, 1512-1522 (published 2023-08-10)
Retrieved: 2026-10-14T12:03:14Z

catalogued

No individual claim review recorded

Audit details

Field: attributes.archive_sha256

Source artifact SHA-256: 919ca366369efbe7e0880565192de127375edddedb1aa43ad023dec612bb4a04

Hash scope: Hash scope not separately documented; inspect source record

Archive member: artifacts/brandes2023-pmc10484790.xml.gz

Inspected artifact

attributes.artifact_sha256
919ca366369efbe7e0880565192de127375edddedb1aa43ad023dec612bb4a04
Source metadata		Genome-wide prediction of disease variant effects with a deep protein language model (Brandes et al., Nature Genetics 2023)

Original source ↗

No field-specific location recorded

Version: Version of record, Nature Genetics 55, 1512-1522 (published 2023-08-10)
Retrieved: 2026-10-14T12:03:14Z

catalogued

No individual claim review recorded

Audit details

Field: attributes.artifact_sha256

Source artifact SHA-256: 919ca366369efbe7e0880565192de127375edddedb1aa43ad023dec612bb4a04

Hash scope: Hash scope not separately documented; inspect source record

Archive member: artifacts/brandes2023-pmc10484790.xml.gz

Inspected artifact

Artifact url
https://www.ebi.ac.uk/europepmc/webservices/rest/PMC10484790/fullTextXML
Source metadata		Genome-wide prediction of disease variant effects with a deep protein language model (Brandes et al., Nature Genetics 2023)

Original source ↗

No field-specific location recorded

Version: Version of record, Nature Genetics 55, 1512-1522 (published 2023-08-10)
Retrieved: 2026-10-14T12:03:14Z

catalogued

No individual claim review recorded

Audit details

Field: attributes.artifact_url

Source artifact SHA-256: 919ca366369efbe7e0880565192de127375edddedb1aa43ad023dec612bb4a04

Hash scope: Hash scope not separately documented; inspect source record

Archive member: artifacts/brandes2023-pmc10484790.xml.gz

Inspected artifact

Doi
10.1038/s41588-023-01465-0
Source metadata		Genome-wide prediction of disease variant effects with a deep protein language model (Brandes et al., Nature Genetics 2023)

Original source ↗

No field-specific location recorded

Version: Version of record, Nature Genetics 55, 1512-1522 (published 2023-08-10)
Retrieved: 2026-10-14T12:03:14Z

catalogued

No individual claim review recorded

Audit details

Field: attributes.doi

Source artifact SHA-256: 919ca366369efbe7e0880565192de127375edddedb1aa43ad023dec612bb4a04

Hash scope: Hash scope not separately documented; inspect source record

Archive member: artifacts/brandes2023-pmc10484790.xml.gz

Inspected artifact

Licence
CC BY 4.0
Source metadata		Genome-wide prediction of disease variant effects with a deep protein language model (Brandes et al., Nature Genetics 2023)

Original source ↗

No field-specific location recorded

Version: Version of record, Nature Genetics 55, 1512-1522 (published 2023-08-10)
Retrieved: 2026-10-14T12:03:14Z

catalogued

No individual claim review recorded

Audit details

Field: attributes.licence

Source artifact SHA-256: 919ca366369efbe7e0880565192de127375edddedb1aa43ad023dec612bb4a04

Hash scope: Hash scope not separately documented; inspect source record

Archive member: artifacts/brandes2023-pmc10484790.xml.gz

Inspected artifact

Limitations
3 values
  • The ROC-AUCs quoted in the Results for ESM1b versus EVE (0.905 and 0.885 on 36,537 ClinVar variants; 0.897 and 0.882 on 30,497 HGMD/gnomAD variants) do not match Supplementary Table 2's EVE head-to-head row (ESM1b 0.906018, EVE 0.885312 on 36,522 ClinVar variants; ESM1b 0.897926, EVE 0.883988 on 30,483 HGMD/gnomAD variants). The text values are extracted as printed; the head-to-head columns are not extracted.
  • Figures 2e-h, 3 and Extended Data figures are not read; the values behind Figure 2e,f are in Supplementary Table 2.
  • HGMD is a licensed resource; the HGMD/gnomAD benchmark cannot be rebuilt from public data.
Source metadata		Genome-wide prediction of disease variant effects with a deep protein language model (Brandes et al., Nature Genetics 2023)

Original source ↗

No field-specific location recorded

Version: Version of record, Nature Genetics 55, 1512-1522 (published 2023-08-10)
Retrieved: 2026-10-14T12:03:14Z

catalogued

No individual claim review recorded

Audit details

Field: attributes.limitations

Source artifact SHA-256: 919ca366369efbe7e0880565192de127375edddedb1aa43ad023dec612bb4a04

Hash scope: Hash scope not separately documented; inspect source record

Archive member: artifacts/brandes2023-pmc10484790.xml.gz

Inspected artifact

Media type
application/xml
Source metadata		Genome-wide prediction of disease variant effects with a deep protein language model (Brandes et al., Nature Genetics 2023)

Original source ↗

No field-specific location recorded

Version: Version of record, Nature Genetics 55, 1512-1522 (published 2023-08-10)
Retrieved: 2026-10-14T12:03:14Z

catalogued

No individual claim review recorded

Audit details

Field: attributes.media_type

Source artifact SHA-256: 919ca366369efbe7e0880565192de127375edddedb1aa43ad023dec612bb4a04

Hash scope: Hash scope not separately documented; inspect source record

Archive member: artifacts/brandes2023-pmc10484790.xml.gz

Inspected artifact

Publication status
peer_reviewed
Source metadata		Genome-wide prediction of disease variant effects with a deep protein language model (Brandes et al., Nature Genetics 2023)

Original source ↗

No field-specific location recorded

Version: Version of record, Nature Genetics 55, 1512-1522 (published 2023-08-10)
Retrieved: 2026-10-14T12:03:14Z

catalogued

No individual claim review recorded

Audit details

Field: attributes.publication_status

Source artifact SHA-256: 919ca366369efbe7e0880565192de127375edddedb1aa43ad023dec612bb4a04

Hash scope: Hash scope not separately documented; inspect source record

Archive member: artifacts/brandes2023-pmc10484790.xml.gz

Inspected artifact

Retrieved at
2026-10-14T12:03:14Z
Source metadata		Genome-wide prediction of disease variant effects with a deep protein language model (Brandes et al., Nature Genetics 2023)

Original source ↗

No field-specific location recorded

Version: Version of record, Nature Genetics 55, 1512-1522 (published 2023-08-10)
Retrieved: 2026-10-14T12:03:14Z

catalogued

No individual claim review recorded

Audit details

Field: attributes.retrieved_at

Source artifact SHA-256: 919ca366369efbe7e0880565192de127375edddedb1aa43ad023dec612bb4a04

Hash scope: Hash scope not separately documented; inspect source record

Archive member: artifacts/brandes2023-pmc10484790.xml.gz

Inspected artifact

How to read the evidence

Sources and history

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

Peer-reviewed article, Europe PMC JATS full text (PMC10484790).

Release history

No supporting source is linked yet.

Read original source

Download this release (gzip)
Technical metadata and extraction receipts

Stable ID: gfm-guides-20261014-esm1b-source-brandes2023-article

areas
dna-genomes; protein-function
contexts
clinical_research
url
https://doi.org/10.1038/s41588-023-01465-0
artifact url
https://www.ebi.ac.uk/europepmc/webservices/rest/PMC10484790/fullTextXML
doi
10.1038/s41588-023-01465-0
version
Version of record, Nature Genetics 55, 1512-1522 (published 2023-08-10)
retrieved at
2026-10-14T12:03:14Z
artifact sha256
919ca366369efbe7e0880565192de127375edddedb1aa43ad023dec612bb4a04
publication status
peer_reviewed
licence
CC BY 4.0
media type
application/xml
venue
Nature Genetics
year
2023
source locator
Results 'ESM1b outperforms other VEP methods over clinical and experimental benchmarks' paragraphs 2 and 5; Methods 'ESM1b', 'AUC metrics for pathogenicity classification', 'Other VEP methods'
archive member
artifacts/brandes2023-pmc10484790.xml.gz
archive sha256
0b3f30de21eff52626d21c249533e7e76e52685d1cfd02cc51b444668b1f58f0
review
method: source-hash-verification; method note: Re-downloaded on 2026-10-14; SHA-256 matches, as does the archived copy.; reviewer: claude; reviewer note: Claude (Opus 5.5) research agent, independent of the extractor; no human review claimed; date: 2026-10-14; artifact sha256: 919ca366369efbe7e0880565192de127375edddedb1aa43ad023dec612bb4a04; retrieval url: https://www.ebi.ac.uk/europepmc/webservices/rest/PMC10484790/fullTextXML; note: Independent review 2026-10-14.
limitations
The ROC-AUCs quoted in the Results for ESM1b versus EVE (0.905 and 0.885 on 36,537 ClinVar variants; 0.897 and 0.882 on 30,497 HGMD/gnomAD variants) do not match Supplementary Table 2's EVE head-to-head row (ESM1b 0.906018, EVE 0.885312 on 36,522 ClinVar variants; ESM1b 0.897926, EVE 0.883988 on 30,483 HGMD/gnomAD variants). The text values are extracted as printed; the head-to-head columns are not extracted.; Figures 2e-h, 3 and Extended Data figures are not read; the values behind Figure 2e,f are in Supplementary Table 2.; HGMD is a licensed resource; the HGMD/gnomAD benchmark cannot be rebuilt from public data.
Related records

    Suggest a correction