ProteinGym ZS-INDEL-AUC: Zero-shot indels, all assays, AUC
Zero-shot indels, all assays, AUC. Scored with AUC on ProteinGym indel DMS assays. Zero-shot scoring of insertion and deletion assays, split by how the test sequences were generated.
Overview
Zero-shot indels, all assays, AUC. Scored with AUC on ProteinGym indel DMS assays. Zero-shot scoring of insertion and deletion assays, split by how the test sequences were generated.
Consult the linked sources for architecture or protocol details. Missing evidence is not evidence of a missing capability.
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.
- HMM on ProteinGym ZS-INDEL-AUC: Zero-shot indels, all assays, AUC
- ProGen2 M on ProteinGym ZS-INDEL-AUC: Zero-shot indels, all assays, AUC
- ProtGPT2 on ProteinGym ZS-INDEL-AUC: Zero-shot indels, all assays, AUC
- PROVEAN on ProteinGym ZS-INDEL-AUC: Zero-shot indels, all assays, AUC
- RITA L on ProteinGym ZS-INDEL-AUC: Zero-shot indels, all assays, AUC
- TranceptEVE M on ProteinGym ZS-INDEL-AUC: Zero-shot indels, all assays, AUC
- Tranception L on ProteinGym ZS-INDEL-AUC: Zero-shot indels, all assays, AUC
- Tranception M on ProteinGym ZS-INDEL-AUC: Zero-shot indels, all assays, AUC
- WaveNet on ProteinGym ZS-INDEL-AUC: Zero-shot indels, all assays, AUC
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.
Published comparisons
Explore the results reported under one evaluation protocol. Each figure keeps its source, dataset and metric together; it is not a ranking across studies. The pooled view gathers every source table that reports the same metric and names what it does not hold constant.
ProteinGym ZS-INDEL-AUC: Zero-shot indels, all assays, AUC
auc (fraction) · Higher values are better for this metric.
Every method ProteinGym reports on Zero-shot indels, all assays, AUC, scored with AUC on ProteinGym indel DMS assays.
Evaluation protocol · ProteinGym indel DMS assays (ProteinGym split)
- HMM · Configuration · Author-reported evaluation0.744
- WaveNet · Configuration · Author-reported evaluation0.720
- PROVEAN · Configuration · Author-reported evaluation0.725
- RITA L · Configuration · Author-reported evaluation0.773
- ProtGPT2 · Configuration · Author-reported evaluation0.620
- ProGen2 M · Configuration · Author-reported evaluation0.776
- Tranception M · Configuration · Author-reported evaluation0.733
- Tranception L · Configuration · Author-reported evaluation0.741
- TranceptEVE M · Configuration · Author-reported evaluation0.754
Source order is preserved. Plotted marks show point estimates; uncertainty, where reported, is retained in the printed values and table. Differences do not establish statistical significance.
ProteinGym: Large-Scale Benchmarks for Protein Fitness Prediction and Design · Table 4, column(Zero-shot indels, all assays, AUC)Values, uncertainty and evidence
| Tested entity | Printed value | Uncertainty | Evidence |
|---|---|---|---|
| HMM · Configuration | 0.744 fraction | Not reported | Author-reported evaluation · source checkedProteinGym: Large-Scale Benchmarks for Protein Fitness Prediction and Design · Table 4, row(HMM), column(Zero-shot indels, all assays, AUC) |
| WaveNet · Configuration | 0.720 fraction | Not reported | Author-reported evaluation · source checkedProteinGym: Large-Scale Benchmarks for Protein Fitness Prediction and Design · Table 4, row(WaveNet), column(Zero-shot indels, all assays, AUC) |
| PROVEAN · Configuration | 0.725 fraction | Not reported | Author-reported evaluation · source checkedProteinGym: Large-Scale Benchmarks for Protein Fitness Prediction and Design · Table 4, row(PROVEAN), column(Zero-shot indels, all assays, AUC) |
| RITA L · Configuration | 0.773 fraction | Not reported | Author-reported evaluation · source checkedProteinGym: Large-Scale Benchmarks for Protein Fitness Prediction and Design · Table 4, row(Protein RITA L), column(Zero-shot indels, all assays, AUC) |
| ProtGPT2 · Configuration | 0.620 fraction | Not reported | Author-reported evaluation · source checkedProteinGym: Large-Scale Benchmarks for Protein Fitness Prediction and Design · Table 4, row(language ProtGPT2), column(Zero-shot indels, all assays, AUC) |
| ProGen2 M · Configuration | 0.776 fraction | Not reported | Author-reported evaluation · source checkedProteinGym: Large-Scale Benchmarks for Protein Fitness Prediction and Design · Table 4, row(models ProGen2 M), column(Zero-shot indels, all assays, AUC) |
| Tranception M · Configuration | 0.733 fraction | Not reported | Author-reported evaluation · source checkedProteinGym: Large-Scale Benchmarks for Protein Fitness Prediction and Design · Table 4, row(Tranception M), column(Zero-shot indels, all assays, AUC) |
| Tranception L · Configuration | 0.741 fraction | Not reported | Author-reported evaluation · source checkedProteinGym: Large-Scale Benchmarks for Protein Fitness Prediction and Design · Table 4, row(Tranception L), column(Zero-shot indels, all assays, AUC) |
| TranceptEVE M · Configuration | 0.754 fraction | Not reported | Author-reported evaluation · source checkedProteinGym: Large-Scale Benchmarks for Protein Fitness Prediction and Design · Table 4, row(TranceptEVE M), column(Zero-shot indels, all assays, AUC) |
Scope and limitations
- 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.
Source transcription and grouping reviewed by automated source review on 2026-09-18. These experiments were not independently reproduced by rewire.
Tested entities and results
Release 2026-09-17-134cd1815de8 · 9 evaluations · 9 metric rows. Different protocols are not a single leaderboard. Where several source tables report the same metric, the published comparisons above offer a pooled view that names what it does not hold constant.
| Metric and finding | Coverage and uncertainty | Evidence |
|---|---|---|
| HMM on ProteinGym ZS-INDEL-AUC: Zero-shot indels, all assays, AUC Configuration: HMMTask: ProteinGym ZS-INDEL-AUC: Zero-shot indels, all assays, AUCDataset subset: ProteinGym indel DMS assays (ProteinGym split) Zero-shot scoring of insertion and deletion assays, split by how the test sequences were generated. Author-reported evaluation · Evaluation metadata: source checked | ||
| 0.744 auc Unit: fraction · Direction: higher | Uncertainty: Not reported Scored: Not reported · Eligible: Not reported | source checkedProteinGym: Large-Scale Benchmarks for Protein Fitness Prediction and Design · Table 4, row(HMM), column(Zero-shot indels, all assays, AUC) Source checking is not independent reproduction. |
| ProGen2 M on ProteinGym ZS-INDEL-AUC: Zero-shot indels, all assays, AUC Configuration: ProGen2 MTask: ProteinGym ZS-INDEL-AUC: Zero-shot indels, all assays, AUCDataset subset: ProteinGym indel DMS assays (ProteinGym split) Zero-shot scoring of insertion and deletion assays, split by how the test sequences were generated. Author-reported evaluation · Evaluation metadata: source checked | ||
| 0.776 auc Unit: fraction · Direction: higher | Uncertainty: Not reported Scored: Not reported · Eligible: Not reported | source checkedProteinGym: Large-Scale Benchmarks for Protein Fitness Prediction and Design · Table 4, row(models ProGen2 M), column(Zero-shot indels, all assays, AUC) Source checking is not independent reproduction. |
| ProtGPT2 on ProteinGym ZS-INDEL-AUC: Zero-shot indels, all assays, AUC Configuration: ProtGPT2Task: ProteinGym ZS-INDEL-AUC: Zero-shot indels, all assays, AUCDataset subset: ProteinGym indel DMS assays (ProteinGym split) Zero-shot scoring of insertion and deletion assays, split by how the test sequences were generated. Author-reported evaluation · Evaluation metadata: source checked | ||
| 0.620 auc Unit: fraction · Direction: higher | Uncertainty: Not reported Scored: Not reported · Eligible: Not reported | source checkedProteinGym: Large-Scale Benchmarks for Protein Fitness Prediction and Design · Table 4, row(language ProtGPT2), column(Zero-shot indels, all assays, AUC) Source checking is not independent reproduction. |
| PROVEAN on ProteinGym ZS-INDEL-AUC: Zero-shot indels, all assays, AUC Configuration: PROVEANTask: ProteinGym ZS-INDEL-AUC: Zero-shot indels, all assays, AUCDataset subset: ProteinGym indel DMS assays (ProteinGym split) Zero-shot scoring of insertion and deletion assays, split by how the test sequences were generated. Author-reported evaluation · Evaluation metadata: source checked | ||
| 0.725 auc Unit: fraction · Direction: higher | Uncertainty: Not reported Scored: Not reported · Eligible: Not reported | source checkedProteinGym: Large-Scale Benchmarks for Protein Fitness Prediction and Design · Table 4, row(PROVEAN), column(Zero-shot indels, all assays, AUC) Source checking is not independent reproduction. |
| RITA L on ProteinGym ZS-INDEL-AUC: Zero-shot indels, all assays, AUC Configuration: RITA LTask: ProteinGym ZS-INDEL-AUC: Zero-shot indels, all assays, AUCDataset subset: ProteinGym indel DMS assays (ProteinGym split) Zero-shot scoring of insertion and deletion assays, split by how the test sequences were generated. Author-reported evaluation · Evaluation metadata: source checked | ||
| 0.773 auc Unit: fraction · Direction: higher | Uncertainty: Not reported Scored: Not reported · Eligible: Not reported | source checkedProteinGym: Large-Scale Benchmarks for Protein Fitness Prediction and Design · Table 4, row(Protein RITA L), column(Zero-shot indels, all assays, AUC) Source checking is not independent reproduction. |
| TranceptEVE M on ProteinGym ZS-INDEL-AUC: Zero-shot indels, all assays, AUC Configuration: TranceptEVE MTask: ProteinGym ZS-INDEL-AUC: Zero-shot indels, all assays, AUCDataset subset: ProteinGym indel DMS assays (ProteinGym split) Zero-shot scoring of insertion and deletion assays, split by how the test sequences were generated. Author-reported evaluation · Evaluation metadata: source checked | ||
| 0.754 auc Unit: fraction · Direction: higher | Uncertainty: Not reported Scored: Not reported · Eligible: Not reported | source checkedProteinGym: Large-Scale Benchmarks for Protein Fitness Prediction and Design · Table 4, row(TranceptEVE M), column(Zero-shot indels, all assays, AUC) Source checking is not independent reproduction. |
| Tranception L on ProteinGym ZS-INDEL-AUC: Zero-shot indels, all assays, AUC Configuration: Tranception LTask: ProteinGym ZS-INDEL-AUC: Zero-shot indels, all assays, AUCDataset subset: ProteinGym indel DMS assays (ProteinGym split) Zero-shot scoring of insertion and deletion assays, split by how the test sequences were generated. Author-reported evaluation · Evaluation metadata: source checked | ||
| 0.741 auc Unit: fraction · Direction: higher | Uncertainty: Not reported Scored: Not reported · Eligible: Not reported | source checkedProteinGym: Large-Scale Benchmarks for Protein Fitness Prediction and Design · Table 4, row(Tranception L), column(Zero-shot indels, all assays, AUC) Source checking is not independent reproduction. |
| Tranception M on ProteinGym ZS-INDEL-AUC: Zero-shot indels, all assays, AUC Configuration: Tranception MTask: ProteinGym ZS-INDEL-AUC: Zero-shot indels, all assays, AUCDataset subset: ProteinGym indel DMS assays (ProteinGym split) Zero-shot scoring of insertion and deletion assays, split by how the test sequences were generated. Author-reported evaluation · Evaluation metadata: source checked | ||
| 0.733 auc Unit: fraction · Direction: higher | Uncertainty: Not reported Scored: Not reported · Eligible: Not reported | source checkedProteinGym: Large-Scale Benchmarks for Protein Fitness Prediction and Design · Table 4, row(Tranception M), column(Zero-shot indels, all assays, AUC) Source checking is not independent reproduction. |
| WaveNet on ProteinGym ZS-INDEL-AUC: Zero-shot indels, all assays, AUC Configuration: WaveNetTask: ProteinGym ZS-INDEL-AUC: Zero-shot indels, all assays, AUCDataset subset: ProteinGym indel DMS assays (ProteinGym split) Zero-shot scoring of insertion and deletion assays, split by how the test sequences were generated. Author-reported evaluation · Evaluation metadata: source checked | ||
| 0.720 auc Unit: fraction · Direction: higher | Uncertainty: Not reported Scored: Not reported · Eligible: Not reported | source checkedProteinGym: Large-Scale Benchmarks for Protein Fitness Prediction and Design · Table 4, row(WaveNet), column(Zero-shot indels, all assays, AUC) Source checking is not 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 4, column(Zero-shot indels, all assays, AUC) 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-indel-auc Source artifact SHA-256: Hash scope: Exact retrieved primary paper artifact bytes. |
Sources and history
Release 2026-09-17-134cd1815de8 · 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-indel-auc
- areas
- proteins-complexes
- tasks
- Zero-shot indels, all assays, AUC
- metric
- AUC
- metric direction
- higher
- dataset
- ProteinGym indel DMS assays
- protocol
- Zero-shot scoring of insertion and deletion assays, split by how the test sequences were generated.
- source locator
- Table 4, column(Zero-shot indels, all assays, AUC)
- comparison panels
- id: proteingym-panel-zs-indel-auc; title: ProteinGym ZS-INDEL-AUC: Zero-shot indels, all assays, AUC; protocol id: proteingym-task-zs-indel-auc; dataset id: proteingym-dataset-proteingym-indel-dms-assays; metric: auc; unit: fraction; direction: higher; result ids: proteingym-result-hmm-zs-indel-auc-auc; proteingym-result-wavenet-zs-indel-auc-auc; proteingym-result-provean-zs-indel-auc-auc; proteingym-result-rita-l-zs-indel-auc-auc; proteingym-result-protgpt2-zs-indel-auc-auc; proteingym-result-progen2-m-zs-indel-auc-auc; proteingym-result-tranception-m-zs-indel-auc-auc; proteingym-result-tranception-l-zs-indel-auc-auc; proteingym-result-trancepteve-m-zs-indel-auc-auc; source ids: evidence-expansion-proteingym-a3b08cc4; source locator: Table 4, column(Zero-shot indels, all assays, AUC); context: Every method ProteinGym reports on Zero-shot indels, all assays, AUC, scored with AUC on ProteinGym indel 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-INDEL-AUC: part of discovery-benchmark-proteingym
- benchmark: HMM on ProteinGym ZS-INDEL-AUC: Zero-shot indels, all assays, AUC
- benchmark: ProGen2 M on ProteinGym ZS-INDEL-AUC: Zero-shot indels, all assays, AUC
- benchmark: ProtGPT2 on ProteinGym ZS-INDEL-AUC: Zero-shot indels, all assays, AUC
- benchmark: PROVEAN on ProteinGym ZS-INDEL-AUC: Zero-shot indels, all assays, AUC
- benchmark: RITA L on ProteinGym ZS-INDEL-AUC: Zero-shot indels, all assays, AUC
- benchmark: TranceptEVE M on ProteinGym ZS-INDEL-AUC: Zero-shot indels, all assays, AUC
- benchmark: Tranception L on ProteinGym ZS-INDEL-AUC: Zero-shot indels, all assays, AUC
- benchmark: Tranception M on ProteinGym ZS-INDEL-AUC: Zero-shot indels, all assays, AUC
- benchmark: WaveNet on ProteinGym ZS-INDEL-AUC: Zero-shot indels, all assays, AUC