ProteinGym Clinical Indels
ProteinGym track for classifying short human clinical insertion and deletion variants against ClinVar and gnomAD-derived labels.
0 evaluation run(s)
suite · audited-with-caveats · verified 2026-07-21
ProteinGym
Versioned deep-mutational-scanning and clinical-variant benchmarks for protein fitness prediction and design in zero-shot and supervised regimes.
+2 more
Audited with caveats: 2 field(s) are marked provisional or conflicted. Warnings are shown next to affected values and these claims are excluded from unqualified comparisons.
Version and binding audit: v1.0 reports 14 generic Binding substitution assays; the v1.1–v1.3 reference files contain 13. The v1.3 release archive contains 66 DMS indel assay records while its pinned README says 74, so the current count remains Conflicted.
Benchmark definition
What is counted
- Version
- 1.3
- Total
- Not reported (heterogeneous DMS assays, clinical proteins, and mutant measurements)
- Task formats
- zero-shot mutation-effect prediction; supervised mutation-effect prediction; clinical variant classification
- Capabilities
- Modalities
Version history
| Version | Status | Release / as-of | Total | Formal tracks |
|---|---|---|---|---|
1.0proteingym-v10 | superseded | 2023-12-08 | Not reported (heterogeneous DMS assays, clinical proteins, and mutant measurements) | proteingym-dms-substitutions, proteingym-dms-indels, proteingym-clinical-substitutions, proteingym-clinical-indels |
1.1proteingym-v11 | superseded | 2024-10-15 | Not reported (heterogeneous DMS assays, clinical proteins, and mutant measurements) | proteingym-dms-substitutions, proteingym-dms-indels, proteingym-clinical-substitutions, proteingym-clinical-indels |
1.2proteingym-v12 | superseded | 2025-03-10 | Not reported (heterogeneous DMS assays, clinical proteins, and mutant measurements) | proteingym-dms-substitutions, proteingym-dms-indels, proteingym-clinical-substitutions, proteingym-clinical-indels |
1.3proteingym-v13 | current | 2025-04-27 | Not reported (heterogeneous DMS assays, clinical proteins, and mutant measurements) | proteingym-dms-substitutions, proteingym-dms-indels, proteingym-clinical-substitutions, proteingym-clinical-indels |
Tracks and subsets
| ID | Count | Basis | Partition? | Notes |
|---|---|---|---|---|
DMS substitution assaysdms-substitution-assays | 217 | assays | No | The v1.3 reference file has 217 assay rows; assay count is distinct from the README description of approximately 2.7 million missense variants. |
DMS indel assaysdms-indel-assays | 66Conflicted · medium | assays | No | The v1.3 versioned archive has 66 assay records; the pinned repository README says 74, so this value is marked Conflicted. |
Clinical substitution proteinsclinical-substitution-proteins | 2525 | proteins | No | The v1.3 clinical reference file has one row per protein. |
Clinical indel proteinsclinical-indel-proteins | 1555 | proteins | No | The v1.3 clinical indel reference file has one row per protein. |
DMS substitution assays in the Binding function categorydms-substitution-binding-assays | 13 | assays | No | The v1.3 reference file contains 13 rows with coarse_selection_type=Binding; target type is not split into protein-protein versus protein-ligand counts. |
Registered child tracks
ProteinGym Clinical Substitutions
ProteinGym track for classifying expert-annotated human clinical substitution variants on a per-protein basis.
0 evaluation run(s)
ProteinGym DMS Indels
ProteinGym track for predicting experimental fitness measurements of insertion and deletion mutants across deep-mutational-scanning assays.
0 evaluation run(s)
ProteinGym DMS Substitutions
ProteinGym track for predicting experimental fitness measurements of substitution mutants across deep-mutational-scanning assays.
1 evaluation run(s)
Scientific Task Atlas
Scientific task classification
partial for 1.3. The release mixes DMS assays, clinical proteins, and mutant measurements; counts stay on formal child tracks.
| Scientific task | Coverage | Count | Mapping | Evidence |
|---|---|---|---|---|
| Protein mutation-effect prediction | explicitly-in-scope | Not reported DMS assays and mutant measurements across version 1.3 tracks. | official-track high confidence | proteingym-evidence-taxonomyTrack-specific assay counts are recorded on child records. |
| Protein fitness prediction | explicitly-in-scope | Not reported ProteinGym DMS assays. | official-taxonomy high confidence | proteingym-evidence-taxonomyNo heterogeneous root total is asserted. |
| Protein clinical variant interpretation | explicitly-in-scope | Not reported Clinical substitution and indel tracks. | official-track high confidence | proteingym-evidence-taxonomyChild records use clinical proteins as their count unit. |
| Protein-ligand binding prediction | observed | Not reported DMS assays in the official generic Binding function category. | official-taxonomy high confidence | proteingym-evidence-taxonomyThe source does not support a ligand-only or PPI-only assay count. |
Scientific coverage notes
| Domain | Coverage | Count | Interpretation |
|---|---|---|---|
| Protein design | explicitly-in-scope | Not reported | NDCG@10% and top-10% recall evaluate design-oriented ranking, but design is not a separately counted task subset. |
| Protein-ligand binding | explicitly-in-scope | Not reported | The paper explicitly includes ligand binding and reports a generic Binding function category: 14 substitution assays in v1.0 and 13 in v1.1-v1.3. It does not publish a target-type split, so no protein-ligand-only count is inferred. |
| Protein-protein binding | observed | Not reported | Some source assays concern protein-protein binding, but the official function category is generic Binding and no protein-protein-only count is published. |
Relationship registry
How works use this benchmark
Partial claims, non-evaluation uses, and third-party summaries stay visible without entering model comparisons.
Partial evaluation claims
evaluation
system-card-claude-opus-5-proteingym-4-use
Work: System Card: Claude Opus 5 · source version system-card-claude-opus-5-2026-07-24
- Selection
- filtered · Subset of mutant protein sequences ranked against the wild type sequence
- Metrics
- rank correlation against real lab measurements
- Linked runs
- None
Not reported / unresolved: Exact ProteinGym version is not reported.; The Hard subset size and selection criteria are not reported.; Prompt, shots, reasoning settings, budget, seed, repeats, grader, and human review are not reported.; The rank-correlation variant and aggregation procedure are not reported.; benchmark version; realized n/scope
AI-assisted double-pass extraction; values are limited to independently supported claims.
Evidence
- section: Section 8.17.3 ProteinGym Hard
Supports: /relation_type - section: Section 8.17.3 ProteinGym Hard
Supports: /benchmark_id - section: Section 8.17.3 ProteinGym Hard
Supports: /scope - section: Section 8.17.3 ProteinGym Hard
Supports: /scope - section: Section 8.17.3 ProteinGym Hard
Supports: /scope - section: Section 8.17.3 ProteinGym Hard
Supports: /metric_labels - section: Section 8.17.3 ProteinGym Hard
Supports: /model_ids - section: Section 8.17.3 ProteinGym Hard
Supports: /model_ids - section: Section 8.17.3 ProteinGym Hard
Supports: /model_ids - section: Section 8.17.3 ProteinGym Hard
Supports: /model_ids
Evaluation registry
Works and run settings
A setting change—scope, prompt, tools, budget, grader, or repeats—creates a separate run. Charts never cross a comparability group.
Evaluation run
proteingym-v10-dms-substitutions-zero-shot
From ProteinGym: Large-Scale Benchmarks for Protein Fitness Prediction and Design
Evaluated models / systems: CARP (38M), CARP (600K), CARP (640M), CARP (76M), DeepSequence (ensemble), DeepSequence (single), ESM-1b, ESM-1v (ensemble), ESM-1v (single), ESM-IF1, ESM2 (150M), ESM2 (15B), ESM2 (35M), ESM2 (3B), ESM2 (650M), ESM2 (8M), EVE (ensemble), EVE (single), EVmutation, GEMME, MIF, MIF-ST, MSA Transformer (ensemble), MSA Transformer (single), ProGen2 Base, ProGen2 L, ProGen2 M, ProGen2 S, ProGen2 XL, ProteinMPNN, ProtGPT2, RITA L, RITA M, RITA S, RITA XL, Site-Independent, TranceptEVE L, TranceptEVE M, TranceptEVE S, Tranception L, Tranception L no retrieval, Tranception M, Tranception M no retrieval, Tranception S, Tranception S no retrieval, UniRep, UniRep evotuned, VESPA, VESPAl, WaveNet
Scopefull · n=217
Shots0
TurnsNot applicable
System prompt publicNot applicable
Reasoning / effortNot applicable
BrowserNot applicable
InternetNot applicable
DatabasesNot applicable
Code executionNot applicable
ContainerNot reported
External toolsmodel-specific inputs
Token budgetNot applicable
Time / cost budgetNot reported
TemperatureNot applicable
SeedNot reported
RepeatsNot reported
Graderdeterministic official scoring pipeline · human review: no
StatisticsNon-parametric bootstrap standard error of each model-to-best Spearman difference over 10,000 bootstrap samples from proteins
ContaminationNot reported
Metrics, results, and full protocol
Metrics
| Metric | Kind / baseline | Unit | Aggregation | Threshold / tolerance |
|---|---|---|---|---|
| Spearman | absolute | correlation | assay-level Spearman, averaged within five functional groups, then equal-weight mean across groups | Not reported |
| AUC | absolute | area under ROC curve | assay-level AUC, averaged within five functional groups, then equal-weight mean across groups | Not reported |
| MCC | absolute | correlation | assay-level MCC, averaged within five functional groups, then equal-weight mean across groups | Not reported |
| NDCG@10% | absolute | normalized discounted cumulative gain | assay-level NDCG@10%, averaged within five functional groups, then equal-weight mean across groups | Not reported |
| Top 10% recall | absolute | proportion | assay-level top-10% recall, averaged within five functional groups, then equal-weight mean across groups | Not reported |
Results
| Model | Metric | Value | n |
|---|---|---|---|
| TranceptEVE L | Spearman | 0.456 correlation Table A5 rank 1*; bootstrap SE of difference from the best model: 0.000. | Not reported |
| TranceptEVE M | Spearman | 0.455 correlation Table A5 rank 1*; bootstrap SE of difference from the best model: 0.004. | Not reported |
| GEMME | Spearman | 0.455 correlation Table A5 rank 1*; bootstrap SE of difference from the best model: 0.007. | Not reported |
| TranceptEVE S | Spearman | 0.452 correlation Table A5 rank 4; bootstrap SE of difference from the best model: 0.004. | Not reported |
| EVE (ensemble) | Spearman | 0.439 correlation Table A5 rank 5; bootstrap SE of difference from the best model: 0.006. | Not reported |
| VESPA | Spearman | 0.436 correlation Table A5 rank 6; bootstrap SE of difference from the best model: 0.006. | Not reported |
| Tranception L | Spearman | 0.434 correlation Table A5 rank 7*; bootstrap SE of difference from the best model: 0.004. | Not reported |
| MSA Transformer (ensemble) | Spearman | 0.434 correlation Table A5 rank 7*; bootstrap SE of difference from the best model: 0.009. | Not reported |
| EVE (single) | Spearman | 0.433 correlation Table A5 rank 9; bootstrap SE of difference from the best model: 0.005. | Not reported |
| Tranception M | Spearman | 0.427 correlation Table A5 rank 10; bootstrap SE of difference from the best model: 0.005. | Not reported |
| ESM-IF1 | Spearman | 0.422 correlation Table A5 rank 11; bootstrap SE of difference from the best model: 0.011. | Not reported |
| MSA Transformer (single) | Spearman | 0.421 correlation Table A5 rank 12; bootstrap SE of difference from the best model: 0.009. | Not reported |
| DeepSequence (ensemble) | Spearman | 0.419 correlation Table A5 rank 13; bootstrap SE of difference from the best model: 0.008. | Not reported |
| Tranception S | Spearman | 0.418 correlation Table A5 rank 14; bootstrap SE of difference from the best model: 0.006. | Not reported |
| ESM2 (650M) | Spearman | 0.414 correlation Table A5 rank 15; bootstrap SE of difference from the best model: 0.012. | Not reported |
| DeepSequence (single) | Spearman | 0.407 correlation Table A5 rank 16*; bootstrap SE of difference from the best model: 0.008. | Not reported |
| ESM-1v (ensemble) | Spearman | 0.407 correlation Table A5 rank 16*; bootstrap SE of difference from the best model: 0.012. | Not reported |
| ESM2 (3B) | Spearman | 0.406 correlation Table A5 rank 18; bootstrap SE of difference from the best model: 0.011. | Not reported |
| MIF-ST | Spearman | 0.401 correlation Table A5 rank 19*; bootstrap SE of difference from the best model: 0.010. | Not reported |
| ESM2 (15B) | Spearman | 0.401 correlation Table A5 rank 19*; bootstrap SE of difference from the best model: 0.010. | Not reported |
| EVmutation | Spearman | 0.395 correlation Table A5 rank 21; bootstrap SE of difference from the best model: 0.006. | Not reported |
| ESM-1b | Spearman | 0.394 correlation Table A5 rank 22*; bootstrap SE of difference from the best model: 0.010. | Not reported |
| VESPAl | Spearman | 0.394 correlation Table A5 rank 22*; bootstrap SE of difference from the best model: 0.007. | Not reported |
| ProGen2 XL | Spearman | 0.391 correlation Table A5 rank 24; bootstrap SE of difference from the best model: 0.008. | Not reported |
| ESM2 (150M) | Spearman | 0.387 correlation Table A5 rank 25; bootstrap SE of difference from the best model: 0.013. | Not reported |
| MIF | Spearman | 0.382 correlation Table A5 rank 26; bootstrap SE of difference from the best model: 0.011. | Not reported |
| ProGen2 L | Spearman | 0.38 correlation Table A5 rank 27; bootstrap SE of difference from the best model: 0.008. | Not reported |
| ProGen2 M | Spearman | 0.379 correlation Table A5 rank 28; bootstrap SE of difference from the best model: 0.008. | Not reported |
| ProGen2 Base | Spearman | 0.378 correlation Table A5 rank 29; bootstrap SE of difference from the best model: 0.009. | Not reported |
| Tranception L no retrieval | Spearman | 0.374 correlation Table A5 rank 30*; bootstrap SE of difference from the best model: 0.008. | Not reported |
| ESM-1v (single) | Spearman | 0.374 correlation Table A5 rank 30*; bootstrap SE of difference from the best model: 0.013. | Not reported |
| WaveNet | Spearman | 0.373 correlation Table A5 rank 32; bootstrap SE of difference from the best model: 0.012. | Not reported |
| RITA XL | Spearman | 0.372 correlation Table A5 rank 33; bootstrap SE of difference from the best model: 0.009. | Not reported |
| CARP (640M) | Spearman | 0.368 correlation Table A5 rank 34; bootstrap SE of difference from the best model: 0.011. | Not reported |
| RITA L | Spearman | 0.365 correlation Table A5 rank 35; bootstrap SE of difference from the best model: 0.009. | Not reported |
| Site-Independent | Spearman | 0.359 correlation Table A5 rank 36; bootstrap SE of difference from the best model: 0.010. | Not reported |
| RITA M | Spearman | 0.35 correlation Table A5 rank 37; bootstrap SE of difference from the best model: 0.010. | Not reported |
| Tranception M no retrieval | Spearman | 0.348 correlation Table A5 rank 38; bootstrap SE of difference from the best model: 0.009. | Not reported |
| UniRep evotuned | Spearman | 0.347 correlation Table A5 rank 39; bootstrap SE of difference from the best model: 0.009. | Not reported |
| ProGen2 S | Spearman | 0.336 correlation Table A5 rank 40; bootstrap SE of difference from the best model: 0.012. | Not reported |
| CARP (76M) | Spearman | 0.328 correlation Table A5 rank 41; bootstrap SE of difference from the best model: 0.012. | Not reported |
| ESM2 (35M) | Spearman | 0.321 correlation Table A5 rank 42; bootstrap SE of difference from the best model: 0.015. | Not reported |
| RITA S | Spearman | 0.304 correlation Table A5 rank 43; bootstrap SE of difference from the best model: 0.011. | Not reported |
| Tranception S no retrieval | Spearman | 0.303 correlation Table A5 rank 44; bootstrap SE of difference from the best model: 0.012. | Not reported |
| CARP (38M) | Spearman | 0.279 correlation Table A5 rank 45; bootstrap SE of difference from the best model: 0.014. | Not reported |
| ProteinMPNN | Spearman | 0.258 correlation Table A5 rank 46; bootstrap SE of difference from the best model: 0.011. | Not reported |
| ESM2 (8M) | Spearman | 0.226 correlation Table A5 rank 47; bootstrap SE of difference from the best model: 0.015. | Not reported |
| UniRep | Spearman | 0.19 correlation Table A5 rank 48; bootstrap SE of difference from the best model: 0.016. | Not reported |
| ProtGPT2 | Spearman | 0.188 correlation Table A5 rank 49; bootstrap SE of difference from the best model: 0.011. | Not reported |
| CARP (600K) | Spearman | 0.106 correlation Table A5 rank 50; bootstrap SE of difference from the best model: 0.016. | Not reported |
Evidence
- section: pp. 5–8 and 29, Table 1, Sections 3.3–4.1, and Table A1 (Defines v1.0, 217 substitution assays, zero-shot label access, baselines, and official scoring.) — supports /scope, /protocol
- table: pp. 3, 6, 9, 35, and 38; Figure 1, Section 4.1, Table 2, Appendix A.5.1, and Table A5 (Defines Spearman, AUC, MCC, NDCG@10%, top-10% recall, corrected-average aggregation, and 10,000-sample bootstrap comparisons.) — supports /metrics, /protocol/statistical
- table: p. 38, Table A5 (All 50 zero-shot DMS substitution Spearman values and bootstrap standard errors of model-to-best differences.) — supports /results
Comparable result views
Spearman
proteingym-v10-dms-substitutions-zero-shot · proteingym-v10-dms-substitutions-zero-shot
Accessible data table
| Model | Value | Comparability group |
|---|---|---|
| TranceptEVE L | 0.456 | proteingym-v10-dms-substitutions-zero-shot |
| TranceptEVE M | 0.455 | proteingym-v10-dms-substitutions-zero-shot |
| GEMME | 0.455 | proteingym-v10-dms-substitutions-zero-shot |
| TranceptEVE S | 0.452 | proteingym-v10-dms-substitutions-zero-shot |
| EVE (ensemble) | 0.439 | proteingym-v10-dms-substitutions-zero-shot |
| VESPA | 0.436 | proteingym-v10-dms-substitutions-zero-shot |
| Tranception L | 0.434 | proteingym-v10-dms-substitutions-zero-shot |
| MSA Transformer (ensemble) | 0.434 | proteingym-v10-dms-substitutions-zero-shot |
| EVE (single) | 0.433 | proteingym-v10-dms-substitutions-zero-shot |
| Tranception M | 0.427 | proteingym-v10-dms-substitutions-zero-shot |
| ESM-IF1 | 0.422 | proteingym-v10-dms-substitutions-zero-shot |
| MSA Transformer (single) | 0.421 | proteingym-v10-dms-substitutions-zero-shot |
| DeepSequence (ensemble) | 0.419 | proteingym-v10-dms-substitutions-zero-shot |
| Tranception S | 0.418 | proteingym-v10-dms-substitutions-zero-shot |
| ESM2 (650M) | 0.414 | proteingym-v10-dms-substitutions-zero-shot |
| DeepSequence (single) | 0.407 | proteingym-v10-dms-substitutions-zero-shot |
| ESM-1v (ensemble) | 0.407 | proteingym-v10-dms-substitutions-zero-shot |
| ESM2 (3B) | 0.406 | proteingym-v10-dms-substitutions-zero-shot |
| MIF-ST | 0.401 | proteingym-v10-dms-substitutions-zero-shot |
| ESM2 (15B) | 0.401 | proteingym-v10-dms-substitutions-zero-shot |
| EVmutation | 0.395 | proteingym-v10-dms-substitutions-zero-shot |
| ESM-1b | 0.394 | proteingym-v10-dms-substitutions-zero-shot |
| VESPAl | 0.394 | proteingym-v10-dms-substitutions-zero-shot |
| ProGen2 XL | 0.391 | proteingym-v10-dms-substitutions-zero-shot |
| ESM2 (150M) | 0.387 | proteingym-v10-dms-substitutions-zero-shot |
| MIF | 0.382 | proteingym-v10-dms-substitutions-zero-shot |
| ProGen2 L | 0.38 | proteingym-v10-dms-substitutions-zero-shot |
| ProGen2 M | 0.379 | proteingym-v10-dms-substitutions-zero-shot |
| ProGen2 Base | 0.378 | proteingym-v10-dms-substitutions-zero-shot |
| Tranception L no retrieval | 0.374 | proteingym-v10-dms-substitutions-zero-shot |
| ESM-1v (single) | 0.374 | proteingym-v10-dms-substitutions-zero-shot |
| WaveNet | 0.373 | proteingym-v10-dms-substitutions-zero-shot |
| RITA XL | 0.372 | proteingym-v10-dms-substitutions-zero-shot |
| CARP (640M) | 0.368 | proteingym-v10-dms-substitutions-zero-shot |
| RITA L | 0.365 | proteingym-v10-dms-substitutions-zero-shot |
| Site-Independent | 0.359 | proteingym-v10-dms-substitutions-zero-shot |
| RITA M | 0.35 | proteingym-v10-dms-substitutions-zero-shot |
| Tranception M no retrieval | 0.348 | proteingym-v10-dms-substitutions-zero-shot |
| UniRep evotuned | 0.347 | proteingym-v10-dms-substitutions-zero-shot |
| ProGen2 S | 0.336 | proteingym-v10-dms-substitutions-zero-shot |
| CARP (76M) | 0.328 | proteingym-v10-dms-substitutions-zero-shot |
| ESM2 (35M) | 0.321 | proteingym-v10-dms-substitutions-zero-shot |
| RITA S | 0.304 | proteingym-v10-dms-substitutions-zero-shot |
| Tranception S no retrieval | 0.303 | proteingym-v10-dms-substitutions-zero-shot |
| CARP (38M) | 0.279 | proteingym-v10-dms-substitutions-zero-shot |
| ProteinMPNN | 0.258 | proteingym-v10-dms-substitutions-zero-shot |
| ESM2 (8M) | 0.226 | proteingym-v10-dms-substitutions-zero-shot |
| UniRep | 0.19 | proteingym-v10-dms-substitutions-zero-shot |
| ProtGPT2 | 0.188 | proteingym-v10-dms-substitutions-zero-shot |
| CARP (600K) | 0.106 | proteingym-v10-dms-substitutions-zero-shot |
Evidence and change history
Source locators remain visible; expand an item to inspect the exact Registry fields it supports.
ProteinGym: Large-Scale Benchmarks for Protein Fitness Prediction and Design · page: pp. 1–6, title, author affiliations, abstract, Figure 1, and Sections 3–4 (Official identity, creator affiliations, benchmark scope, modalities, task formats, and evaluation regimes.) · Supports 9 fields
/name/organizations/release_date/kind/summary/domains/capabilities/modalities/task_formats
ProteinGym: Large-Scale Benchmarks for Protein Fitness Prediction and Design · table: pp. 5–6 and 30, Table 1, Section 4.1, and Table A2 (DMS/clinical regimes, design-oriented metrics, ligand-binding scope, and the generic Binding function category.) · Supports 9 fields
/domains/capabilities/modalities/task_formats/coverage_notes/scientific_task_classification/entries/0/scientific_task_classification/entries/1/scientific_task_classification/entries/2/scientific_task_classification/entries/3
ProteinGym: Large-Scale Benchmarks for Protein Fitness Prediction and Design · table: pp. 6, 29–30; Tables 1, A1, and A2 (v1.0 reports 217 substitution assays, 66 indel assays, 2,525/1,555 clinical proteins, and 14 generic Binding substitution assays.) · Supports 3 fields
/versions/0/task_counts/total/versions/0/task_counts/basis/versions/0/task_counts/subsets
proteingym-v10-dataset-resource · dataset-card: Zenodo record 13932633 metadata: version 1.0, publication date 2023-12, open access, MIT license (The benchmark release date is anchored to the creator preprint posted on 2023-12-08 because the dataset record only reports month precision.) · Supports 3 fields
/release_date/versions/0/release_date/resources
proteingym-v11-dataset-resource · repository-path: ProteinGym_v1.1.zip: DMS_substitutions.csv, DMS_indels.csv, clinical_substitutions.csv, clinical_indels.csv (CSV row counts are 217, 66, 2,525, and 1,555; coarse_selection_type has 13 Binding rows.) · Supports 3 fields
/versions/1/task_counts/total/versions/1/task_counts/basis/versions/1/task_counts/subsets
proteingym-v12-dataset-resource · repository-path: Zenodo record 14997691: DMS_substitutions.csv, DMS_indels.csv, clinical_substitutions.csv, clinical_indels.csv (CSV row counts are 217, 66, 2,525, and 1,555; coarse_selection_type has 13 Binding rows.) · Supports 3 fields
/versions/2/task_counts/total/versions/2/task_counts/basis/versions/2/task_counts/subsets
proteingym-v13-dataset-resource · repository-path: Zenodo record 15293562: DMS_substitutions.csv, DMS_indels.csv, clinical_substitutions.csv, clinical_indels.csv and DMS_ProteinGym_indels.zip (Reference/archive row counts are 217, 66, 2,525, and 1,555; coarse_selection_type has 13 Binding rows.) · Supports 8 fields
/task_counts/total/task_counts/basis/task_counts/subsets/task_counts/subsets/1/count/versions/3/task_counts/total/versions/3/task_counts/basis/versions/3/task_counts/subsets/versions/3/task_counts/subsets/1/count
proteingym-repository-resource · release: PG_v1.0–PG_v1.3 tags and README Releases section at commit 1f8de974dead8ff7501eff087b725d14a965e9f9 (The README identifies v1.3 as latest and links all four immutable Zenodo releases.) · Supports 6 fields
/latest_version/versions/1/release_date/versions/2/release_date/versions/3/release_date/resources/implementations
proteingym-repository-resource · repository-path: README.md, Overview and Results at tag PG_v1.3 (README reports 217 DMS substitution assays, 74 DMS indel assays, 2,525/1,555 clinical proteins, metrics, aggregation, downloads, and MIT license.) · Supports 9 fields
/task_counts/subsets/1/count/versions/3/task_counts/subsets/1/count/access/level/access/tasks/access/artifacts/access/grader/access/license/resources/implementations
ProteinGym: Large-Scale Benchmarks for Protein Fitness Prediction and Design · section: pp. 11 and 28–29, Section 6 and Appendices A.1 and A.3.3–A.3.4 (Public code/data/model-score resources, MIT code license, and explicit dual-use discussion.) · Supports 8 fields
/access/level/access/tasks/access/artifacts/access/grader/access/license/access/biosafety_notes/resources/implementations
Unresolved field claims
/task_counts/subsets/1/countConflicted · medium — The versioned v1.3 archive contains 66 DMS indel assay records, while the official PG_v1.3 README states 74. The archive value is retained under the source-priority policy.
Evidence: proteingym-evidence-v13-counts, proteingym-evidence-current-readme/versions/3/task_counts/subsets/1/countConflicted · medium — The versioned v1.3 archive contains 66 DMS indel assay records, while the official PG_v1.3 README states 74. The archive value is retained under the source-priority policy.
Evidence: proteingym-evidence-v13-counts, proteingym-evidence-current-readme