ProteinGym ZS-SUB-RECALL: Zero-shot substitutions, top 10% recall
Zero-shot substitutions, top 10% recall. Scored with Top 10% recall on ProteinGym substitution DMS assays. Zero-shot scoring of substitution assays, averaged over assays with the correction the paper describes.
Overview
Zero-shot substitutions, top 10% recall. Scored with Top 10% recall on ProteinGym substitution DMS assays. Zero-shot scoring of substitution assays, averaged over assays with the correction the paper describes.
Consult the linked sources for architecture or protocol details. Missing evidence is not evidence of a missing capability.
Results
Each comparison retains its reviewed evaluation scope, dataset and metric. Results are shown without a pooled ranking.
ProteinGym ZS-SUB-RECALL: Zero-shot substitutions, top 10% recall
recall (fraction) · Higher values are better.
ProteinGym ZS-SUB-RECALL: Zero-shot substitutions, top 10% recall · ProteinGym substitution DMS assays (ProteinGym split)
Evidence origin: Author-reported evaluation.
ProteinGym: Large-Scale Benchmarks for Protein Fitness Prediction and Design · Table 2, column(Zero-shot substitutions, top 10% recall)- MSE is better when lower; every other metric here is better when higher.
- Zero-shot, supervised and indel benchmarks use different protocols and assay sets, so their figures are not comparable to each other.
Comparison details and limitations
Every method ProteinGym reports on Zero-shot substitutions, top 10% recall, scored with Top 10% recall on ProteinGym substitution DMS assays.
- Author-reported numbers, source checked but not independently reproduced.
Automated source review: 2026-09-18. Numerical source review does not establish independent reproduction.
Dots show point estimates. Whiskers show only explicitly defined uncertainty (standard deviation, standard error or a labelled interval); their definitions remain in Table. Unresolved uncertainty is not plotted. Differences do not establish statistical significance.
Showing 12 of 21 matching rows.
Methods and evaluation design
Procedure, tasks and evaluated configurations
Evaluation design
Benchmarks bring together tasks and protocols. A task describes the biological question; a protocol defines a particular test.
Benchmarks
These source-backed links do not make different protocols or scores interchangeable.
Recorded evaluations
Each evaluation records what was tested and under which conditions.
- CARP (640M) on ProteinGym ZS-SUB-RECALL: Zero-shot substitutions, top 10% recall
- DeepSequence (ensemble) on ProteinGym ZS-SUB-RECALL: Zero-shot substitutions, top 10% recall
- ESM-1b on ProteinGym ZS-SUB-RECALL: Zero-shot substitutions, top 10% recall
- ESM-1v (ensemble) on ProteinGym ZS-SUB-RECALL: Zero-shot substitutions, top 10% recall
- ESM-2 (15B) on ProteinGym ZS-SUB-RECALL: Zero-shot substitutions, top 10% recall
- ESM-IF1 on ProteinGym ZS-SUB-RECALL: Zero-shot substitutions, top 10% recall
- EVE (ensemble) on ProteinGym ZS-SUB-RECALL: Zero-shot substitutions, top 10% recall
- EVmutation on ProteinGym ZS-SUB-RECALL: Zero-shot substitutions, top 10% recall
- GEMME on ProteinGym ZS-SUB-RECALL: Zero-shot substitutions, top 10% recall
- MIF-ST on ProteinGym ZS-SUB-RECALL: Zero-shot substitutions, top 10% recall
- MSA Transformer (ensemble) on ProteinGym ZS-SUB-RECALL: Zero-shot substitutions, top 10% recall
- ProGen2 XL on ProteinGym ZS-SUB-RECALL: Zero-shot substitutions, top 10% recall
Run instructions
No runnable recipe has been reviewed for this task. Dataset access, model requirements, licences and compute requirements must be checked against its sources before execution.
A task describes a biological question. Choose a linked protocol to obtain concrete split and scoring instructions.
Strengths, limitations and unresolved questions
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
| Property and statement | Original source and location | Review and provenance |
|---|---|---|
| Relationship: part of discovery-benchmark-proteingym Individual claims | ProteinGym: Large-Scale Benchmarks for Protein Fitness Prediction and Design Table 2, column(Zero-shot substitutions, top 10% recall) Version: Primary full-text snapshot retrieved 2026-09-17; exact bytes pinned by SHA-256 | source checked automated source review · 2026-09-18 Audit detailsPrimary-source transcription with no human sign-off and no independent reproduction. Field: Claim: proteingym-association-zs-sub-recall Source artifact SHA-256: Hash scope: Exact retrieved primary paper artifact bytes. |
Sources and history
View linked audit checks and correction history
Release 2026-09-29-06401fd5b220 · Record review: source checked
1 source records and release history
- ProteinGym: Large-Scale Benchmarks for Protein Fitness Prediction and Design · Original source · Primary full-text snapshot retrieved 2026-09-17; exact bytes pinned by SHA-256
Technical metadata and extraction receipts
Stable ID: proteingym-task-zs-sub-recall
- areas
- proteins-complexes
- tasks
- Zero-shot substitutions, top 10% recall
- metric
- Top 10% recall
- metric direction
- higher
- dataset
- ProteinGym substitution DMS assays
- protocol
- Zero-shot scoring of substitution assays, averaged over assays with the correction the paper describes.
- source locator
- Table 2, column(Zero-shot substitutions, top 10% recall)
- comparison panels
- id: proteingym-panel-zs-sub-recall; title: ProteinGym ZS-SUB-RECALL: Zero-shot substitutions, top 10% recall; protocol id: proteingym-task-zs-sub-recall; dataset id: proteingym-dataset-proteingym-substitution-dms-assays; metric: recall; unit: fraction; direction: higher; result ids: proteingym-result-site-independent-zs-sub-recall-recall; proteingym-result-wavenet-zs-sub-recall-recall; proteingym-result-evmutation-zs-sub-recall-recall; proteingym-result-deepsequence-ensemble-zs-sub-recall-recall; proteingym-result-eve-ensemble-zs-sub-recall-recall; proteingym-result-gemme-zs-sub-recall-recall; proteingym-result-unirep-zs-sub-recall-recall; proteingym-result-carp-640m-zs-sub-recall-recall; proteingym-result-esm-1b-zs-sub-recall-recall; proteingym-result-esm-2-15b-zs-sub-recall-recall; proteingym-result-rita-xl-zs-sub-recall-recall; proteingym-result-esm-1v-ensemble-zs-sub-recall-recall; proteingym-result-progen2-xl-zs-sub-recall-recall; proteingym-result-vespa-zs-sub-recall-recall; proteingym-result-unirep-evotuned-zs-sub-recall-recall; proteingym-result-msa-transformer-ensemble-zs-sub-recall-recall; proteingym-result-tranception-l-zs-sub-recall-recall; proteingym-result-trancepteve-l-zs-sub-recall-recall; proteingym-result-esm-if1-zs-sub-recall-recall; proteingym-result-mif-st-zs-sub-recall-recall; proteingym-result-proteinmpnn-zs-sub-recall-recall; source ids: evidence-expansion-proteingym-a3b08cc4; source locator: Table 2, column(Zero-shot substitutions, top 10% recall); context: Every method ProteinGym reports on Zero-shot substitutions, top 10% recall, scored with Top 10% recall on ProteinGym substitution DMS assays.; caveats: Author-reported numbers, source checked but not independently reproduced.; MSE is better when lower; every other metric here is better when higher.; Zero-shot, supervised and indel benchmarks use different protocols and assay sets, so their figures are not comparable to each other.; review: method: automated_source_review; date: 2026-09-18
Related records
- part of: ProteinGym
- subject: ProteinGym ZS-SUB-RECALL: part of discovery-benchmark-proteingym
- benchmark: CARP (640M) on ProteinGym ZS-SUB-RECALL: Zero-shot substitutions, top 10% recall
- benchmark: DeepSequence (ensemble) on ProteinGym ZS-SUB-RECALL: Zero-shot substitutions, top 10% recall
- benchmark: ESM-1b on ProteinGym ZS-SUB-RECALL: Zero-shot substitutions, top 10% recall
- benchmark: ESM-1v (ensemble) on ProteinGym ZS-SUB-RECALL: Zero-shot substitutions, top 10% recall
- benchmark: ESM-2 (15B) on ProteinGym ZS-SUB-RECALL: Zero-shot substitutions, top 10% recall
- benchmark: ESM-IF1 on ProteinGym ZS-SUB-RECALL: Zero-shot substitutions, top 10% recall
- benchmark: EVE (ensemble) on ProteinGym ZS-SUB-RECALL: Zero-shot substitutions, top 10% recall
- benchmark: EVmutation on ProteinGym ZS-SUB-RECALL: Zero-shot substitutions, top 10% recall
- benchmark: GEMME on ProteinGym ZS-SUB-RECALL: Zero-shot substitutions, top 10% recall
- benchmark: MIF-ST on ProteinGym ZS-SUB-RECALL: Zero-shot substitutions, top 10% recall
- benchmark: MSA Transformer (ensemble) on ProteinGym ZS-SUB-RECALL: Zero-shot substitutions, top 10% recall
- benchmark: ProGen2 XL on ProteinGym ZS-SUB-RECALL: Zero-shot substitutions, top 10% recall
- benchmark: ProteinMPNN on ProteinGym ZS-SUB-RECALL: Zero-shot substitutions, top 10% recall
- benchmark: RITA XL on ProteinGym ZS-SUB-RECALL: Zero-shot substitutions, top 10% recall
- benchmark: Site-Independent on ProteinGym ZS-SUB-RECALL: Zero-shot substitutions, top 10% recall
- benchmark: TranceptEVE L on ProteinGym ZS-SUB-RECALL: Zero-shot substitutions, top 10% recall
- benchmark: Tranception L on ProteinGym ZS-SUB-RECALL: Zero-shot substitutions, top 10% recall
- benchmark: UniRep evotuned on ProteinGym ZS-SUB-RECALL: Zero-shot substitutions, top 10% recall
- benchmark: UniRep on ProteinGym ZS-SUB-RECALL: Zero-shot substitutions, top 10% recall
- benchmark: VESPA on ProteinGym ZS-SUB-RECALL: Zero-shot substitutions, top 10% recall
- benchmark: WaveNet on ProteinGym ZS-SUB-RECALL: Zero-shot substitutions, top 10% recall