ESM-2 8M masked-marginal scoring on ProteinGym v1.3 AMFR substitution assay
Score all 2,972 variants in AMFR_HUMAN_Tsuboyama_2023_4G3O with ESM-2 8M masked marginals. This single-assay result is not a full ProteinGym track or suite score.
Evaluation results
1 evaluation · 5 metric rows. Different protocols are not a single leaderboard.
Filter evaluations
Applied filters: All linked evaluations
Source checking is not independent reproduction. Release 2026-09-22-f58a0f1d267f.
Evaluation procedure
ProteinGym v1.3 AMFR substitution assay
- Configuration
- ESM-2 8M masked-marginal scoring (ProteinGym v1.3 AMFR substitution assay)
- Protocol
- ProteinGym v1.3 AMFR substitution assay
- Dataset subset
- ProteinGym v1.3 AMFR substitution assay full assay
- origin
- Rewire evaluation
- configuration
- 1.3
- protocol id
- rewire-protocol-proteingym-amfr-v13
- dataset version
- rewire-dataset-proteingym-amfr-v13
- split
- Full assay; no train/test split
- subset
- proteingym-dms-substitutions-v1.3
- population
- 2972/2972
- inputs
- Wild-type protein sequence and amino-acid substitutions; no MSA, structure or assay labels
- adaptation
- zero-shot masked marginals
- metric implementation
- fcc5ab65e6753cf29aabbdae231cbb5cdf72a3dd40ba8f109802c33223b33536
- aggregation
- single assay
- budget
- one local CPU evaluation; no hyperparameter search outside training
Metadata review: source checked. Unreported conditions prevent automatic comparisons.
Reproduction
- Split
- Full assay; no train/test split
- Adaptation
- zero-shot masked marginals
- Scoring implementation
- fcc5ab65e6753cf29aabbdae231cbb5cdf72a3dd40ba8f109802c33223b33536
Generate and evaluate predictions
This committed script regenerates the selected AMFR assay using the recorded checkpoint and masked-marginal scorer. It is not the complete ProteinGym track or a reproduction of a paper score.
ProteinGym v1.3 AMFR substitution assay: reproduction instructions · Reproduce sectionReproducing 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.
152 evidence rows matching the loaded filters
| Property and statement | Original source and location | Review and provenance |
|---|---|---|
| attributes.comparison.adaptation zero-shot masked marginals Context-only references | ProteinGym v1.3 AMFR substitution assay: reproduction instructions metrics; coverage; protocol_results Shared locator for this statement’s cited sources; not a separate locator for each citation. Version: ca73fa47136d182f2d4ddb083d084712198fc0e2 | not individually reviewed No individual claim review recorded rewire run Audit detailsField: Source artifact SHA-256: Hash scope: Hash scope not separately documented; inspect source record |
| attributes.comparison.adaptation zero-shot masked marginals Context-only references | ESM-2 8M masked-marginal scoring: local execution report (20 September 2026) metrics; coverage; protocol_results Shared locator for this statement’s cited sources; not a separate locator for each citation. Version: ca73fa47136d182f2d4ddb083d084712198fc0e2 | not individually reviewed No individual claim review recorded rewire run Audit detailsField: Source artifact SHA-256: Hash scope: Hash scope not separately documented; inspect source record |
| attributes.comparison.aggregation single assay Context-only references | ProteinGym v1.3 AMFR substitution assay: reproduction instructions metrics; coverage; protocol_results Shared locator for this statement’s cited sources; not a separate locator for each citation. Version: ca73fa47136d182f2d4ddb083d084712198fc0e2 | not individually reviewed No individual claim review recorded rewire run Audit detailsField: Source artifact SHA-256: Hash scope: Hash scope not separately documented; inspect source record |
| attributes.comparison.aggregation single assay Context-only references | ESM-2 8M masked-marginal scoring: local execution report (20 September 2026) metrics; coverage; protocol_results Shared locator for this statement’s cited sources; not a separate locator for each citation. Version: ca73fa47136d182f2d4ddb083d084712198fc0e2 | not individually reviewed No individual claim review recorded rewire run Audit detailsField: Source artifact SHA-256: Hash scope: Hash scope not separately documented; inspect source record |
| attributes.comparison.budget one local CPU evaluation; no hyperparameter search outside training Context-only references | ProteinGym v1.3 AMFR substitution assay: reproduction instructions metrics; coverage; protocol_results Shared locator for this statement’s cited sources; not a separate locator for each citation. Version: ca73fa47136d182f2d4ddb083d084712198fc0e2 | not individually reviewed No individual claim review recorded rewire run Audit detailsField: Source artifact SHA-256: Hash scope: Hash scope not separately documented; inspect source record |
| attributes.comparison.budget one local CPU evaluation; no hyperparameter search outside training Context-only references | ESM-2 8M masked-marginal scoring: local execution report (20 September 2026) metrics; coverage; protocol_results Shared locator for this statement’s cited sources; not a separate locator for each citation. Version: ca73fa47136d182f2d4ddb083d084712198fc0e2 | not individually reviewed No individual claim review recorded rewire run Audit detailsField: Source artifact SHA-256: Hash scope: Hash scope not separately documented; inspect source record |
| attributes.comparison.dataset_version rewire-dataset-proteingym-amfr-v13 Context-only references | ProteinGym v1.3 AMFR substitution assay: reproduction instructions metrics; coverage; protocol_results Shared locator for this statement’s cited sources; not a separate locator for each citation. Version: ca73fa47136d182f2d4ddb083d084712198fc0e2 | not individually reviewed No individual claim review recorded rewire run Audit detailsField: Source artifact SHA-256: Hash scope: Hash scope not separately documented; inspect source record |
| attributes.comparison.dataset_version rewire-dataset-proteingym-amfr-v13 Context-only references | ESM-2 8M masked-marginal scoring: local execution report (20 September 2026) metrics; coverage; protocol_results Shared locator for this statement’s cited sources; not a separate locator for each citation. Version: ca73fa47136d182f2d4ddb083d084712198fc0e2 | not individually reviewed No individual claim review recorded rewire run Audit detailsField: Source artifact SHA-256: Hash scope: Hash scope not separately documented; inspect source record |
| attributes.comparison.inputs Wild-type protein sequence and amino-acid substitutions; no MSA, structure or assay labels Context-only references | ProteinGym v1.3 AMFR substitution assay: reproduction instructions metrics; coverage; protocol_results Shared locator for this statement’s cited sources; not a separate locator for each citation. Version: ca73fa47136d182f2d4ddb083d084712198fc0e2 | not individually reviewed No individual claim review recorded rewire run Audit detailsField: Source artifact SHA-256: Hash scope: Hash scope not separately documented; inspect source record |
| attributes.comparison.inputs Wild-type protein sequence and amino-acid substitutions; no MSA, structure or assay labels Context-only references | ESM-2 8M masked-marginal scoring: local execution report (20 September 2026) metrics; coverage; protocol_results Shared locator for this statement’s cited sources; not a separate locator for each citation. Version: ca73fa47136d182f2d4ddb083d084712198fc0e2 | not individually reviewed No individual claim review recorded rewire run Audit detailsField: Source artifact SHA-256: Hash scope: Hash scope not separately documented; inspect source record |
Sources and history
View linked audit checks and correction history
Release 2026-09-22-f58a0f1d267f · Record review: source checked
2 source records and release history
- ESM-2 8M masked-marginal scoring: local execution report (20 September 2026) · Original source · ca73fa47136d182f2d4ddb083d084712198fc0e2
- ProteinGym v1.3 AMFR substitution assay: reproduction instructions · Original source · ca73fa47136d182f2d4ddb083d084712198fc0e2
Technical metadata and extraction receipts
Stable ID: rewire-local-20260920-evaluation-proteingym-esm2
- areas
- proteins-complexes
- origin
- rewire_run
- execution scope
- complete_selected_evaluation
- published score reproduction
- false
- protocol
- ProteinGym v1.3 AMFR substitution assay
- version
- 1.3
- source locator
- metrics; coverage; protocol_results
- original sdk scope
- subset
- original sdk completion
- partial
- eligible count
- 2972
- scored count
- 2972
- missing count
- 0
- coverage
- 2972/2972
- suite complete
- false
- comparison
- protocol id: rewire-protocol-proteingym-amfr-v13; dataset version: rewire-dataset-proteingym-amfr-v13; split: Full assay; no train/test split; subset: proteingym-dms-substitutions-v1.3; population: 2972/2972; inputs: Wild-type protein sequence and amino-acid substitutions; no MSA, structure or assay labels; adaptation: zero-shot masked marginals; metric implementation: fcc5ab65e6753cf29aabbdae231cbb5cdf72a3dd40ba8f109802c33223b33536; aggregation: single assay; budget: one local CPU evaluation; no hyperparameter search outside training
- input information
- Wild-type protein sequence and amino-acid substitutions; no MSA, structure or assay labels
- model configuration
- checkpoint: esm2_t6_8M_UR50D; device: cpu; seed: 0; strategy: masked_marginals; threads: 1
- provenance
- assay sha256: AMFR HUMAN Tsuboyama 2023 4G3O: dd911d925eca79329a496eb6cb7b181ab448e15db244daf1c1f9c01e8a704c2a; data verification: local_bytes_hashed_not_independently_source_verified; expected hashes checked: AMFR_HUMAN_Tsuboyama_2023_4G3O; reference sha256: a8f498011532a74aa9fe556a50555a75e928c5837d19c06a87592ae04049b308; source url: https://github.com/OATML-Markslab/ProteinGym/tree/144fe22b07dfaeec2b366f2346203a9838a55b4c; upstream revision: 144fe22b07dfaeec2b366f2346203a9838a55b4c
- predictions sha256
- 7ab0c8901406d5c1ab759038c8beef235721938394cfdf192c46671e98e0bbca
- prepared sha256
- 9f0e0cce34942797364990ef4dc9fb8b382f7b2fd59e6c72de23efc2992fb148
- execution
- adapter: ESM2Adapter; adapter provenance: checkpoint sha256: 46f002a9870c9bdecd0ea887acb1f9a38a6b561e8f8bf8a6990b679b9d31b928; checkpoint url: https://dl.fbaipublicfiles.com/fair-esm/models/esm2_t6_8M_UR50D.pt; device: cpu; fair esm version: 2.0.0; implementation sha256: 4ff38cfbb91d718b52a44da53aa5594fe73d9d1d4d0ce43724807597c3929889; maximum sequence length: 1022; model: esm2_t6_8M_UR50D; published result reproduction: false; strategy: masked_marginals_wild_type_context_sum_log_odds; training overlap: unreported; UniRef50 pretraining may overlap benchmark proteins; batch size: 32; fitting: none; inference and fit seconds: 0.18779424999956973; mode: local_adapter; prediction type: scalar
- timing seconds
- evaluation: 0.03425099999731174
- environment
- architecture: arm64; container digest: unreported; container runtime: unreported; platform: macOS-26.6.2-arm64-arm-64bit; python: 3.11.13; rewirebench: 0.4.0; sdk code sha256: fcc5ab65e6753cf29aabbdae231cbb5cdf72a3dd40ba8f109802c33223b33536; sif sha256: unreported
- run url
- https://github.com/rewire-bio/rewire-benchmarks/blob/ca73fa47136d182f2d4ddb083d084712198fc0e2/research/local-runs-2026-09-20/proteingym-esm2/report.json
- reproduction url
- https://github.com/rewire-bio/rewire-benchmarks/blob/ca73fa47136d182f2d4ddb083d084712198fc0e2/research/local-runs-2026-09-20/README.md
- limitations
- One selected split or assay; no whole-suite score.; No published model score reproduction claim.; Raw predictions retained locally; hashes and inspected contribution bundle are public.; Single execution; no seed variability or uncertainty interval estimated.
- reproduction
- recipe owner id: rewire-protocol-proteingym-amfr-v13; recipe id: local-20260920-proteingym-repeat; applicability: generate_and_evaluate; explanation: This committed script regenerates the selected AMFR assay using the recorded checkpoint and masked-marginal scorer. It is not the complete ProteinGym track or a reproduction of a paper score.; source ids: rewire-local-20260920-instructions-proteingym; source locator: Reproduce section
Related records
- configuration: ESM-2 8M masked-marginal scoring (ProteinGym v1.3 AMFR substitution assay)
- protocol: ProteinGym v1.3 AMFR substitution assay
- dataset subset: ProteinGym v1.3 AMFR substitution assay full assay
- evaluation: ESM-2 8M masked-marginal scoring: AUC
- evaluation: ESM-2 8M masked-marginal scoring: MCC
- evaluation: ESM-2 8M masked-marginal scoring: NDCG
- evaluation: ESM-2 8M masked-marginal scoring: Spearman
- evaluation: ESM-2 8M masked-marginal scoring: Top_recall