FLIP RANDOM-SAMPLED-SPLITS-GB1: Random sampled splits fitness prediction, GB1 split
Random sampled splits fitness prediction, GB1 split. Scored with Spearman correlation on FLIP Random sampled splits, GB1 split. Trained on the split's training set with ten percent held out for validation, scored on its test set. The authors report these separately and call them optimistic: a random split of a mutational landscape leaks close relatives between train and test.
Overview
Random sampled splits fitness prediction, GB1 split. Scored with Spearman correlation on FLIP Random sampled splits, GB1 split. Trained on the split's training set with ten percent held out for validation, scored on its test set. The authors report these separately and call them optimistic: a random split of a mutational landscape leaks close relatives between train and test.
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.
- CNN on FLIP RANDOM-SAMPLED-SPLITS-GB1: Random sampled splits fitness prediction, GB1 split
- ESM-1b (per AA) on FLIP RANDOM-SAMPLED-SPLITS-GB1: Random sampled splits fitness prediction, GB1 split
- ESM-1v (per AA) on FLIP RANDOM-SAMPLED-SPLITS-GB1: Random sampled splits fitness prediction, GB1 split
- ESM-untrained (per AA) on FLIP RANDOM-SAMPLED-SPLITS-GB1: Random sampled splits fitness prediction, GB1 split
- Ridge on FLIP RANDOM-SAMPLED-SPLITS-GB1: Random sampled splits fitness prediction, GB1 split
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.
FLIP RANDOM-SAMPLED-SPLITS-GB1: Random sampled splits fitness prediction, GB1 split
spearman (correlation) · Higher values are better for this metric.
Every method FLIP reports on Random sampled splits fitness prediction, GB1 split, scored with Spearman correlation on FLIP Random sampled splits, GB1 split.
Evaluation protocol · FLIP Random sampled splits, GB1 split (FLIP split)
- ESM-1b (per AA) · Configuration · Author-reported evaluation0.92
- ESM-1v (per AA) · Configuration · Author-reported evaluation0.92
- ESM-untrained (per AA) · Configuration · Author-reported evaluation0.79
- Ridge · Method · Author-reported evaluation0.82
- CNN · Method · Author-reported evaluation0.91
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.
flip primary benchmark evidence · Table 7, column(GB1)Values, uncertainty and evidence
| Tested entity | Printed value | Uncertainty | Evidence |
|---|---|---|---|
| ESM-1b (per AA) · Configuration | 0.92 correlation | Not reported | Author-reported evaluation · source checkedflip primary benchmark evidence · Table 7, row(ESM-1b (per AA)), column(GB1) |
| ESM-1v (per AA) · Configuration | 0.92 correlation | Not reported | Author-reported evaluation · source checkedflip primary benchmark evidence · Table 7, row(ESM-1v (per AA)), column(GB1) |
| ESM-untrained (per AA) · Configuration | 0.79 correlation | Not reported | Author-reported evaluation · source checkedflip primary benchmark evidence · Table 7, row(ESM-untrained (per AA)), column(GB1) |
| Ridge · Method | 0.82 correlation | Not reported | Author-reported evaluation · source checkedflip primary benchmark evidence · Table 7, row(Ridge), column(GB1) |
| CNN · Method | 0.91 correlation | Not reported | Author-reported evaluation · source checkedflip primary benchmark evidence · Table 7, row(CNN), column(GB1) |
Scope and limitations
- Author-reported numbers, source checked but not independently reproduced.
- The split is the task here. Figures from different splits of the same landscape are not comparable.
- The random sampled splits in Table 7 are the authors' own illustration of an optimistic evaluation, not a headline result.
- NA marks a baseline that cannot be applied to that landscape, and is recorded as no result rather than as a zero.
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 · 5 evaluations · 5 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 |
|---|---|---|
| CNN on FLIP RANDOM-SAMPLED-SPLITS-GB1: Random sampled splits fitness prediction, GB1 split Method: CNNTask: FLIP RANDOM-SAMPLED-SPLITS-GB1: Random sampled splits fitness prediction, GB1 splitDataset subset: FLIP Random sampled splits, GB1 split (FLIP split) Trained on the split's training set with ten percent held out for validation, scored on its test set. The authors report these separately and call them optimistic: a random split of a mutational landscape leaks close relatives between train and test. Author-reported evaluation · Evaluation metadata: source checked | ||
| 0.91 spearman Unit: correlation · Direction: higher | Uncertainty: Not reported Scored: Not reported · Eligible: Not reported | source checkedflip primary benchmark evidence · Table 7, row(CNN), column(GB1) Source checking is not independent reproduction. |
| ESM-1b (per AA) on FLIP RANDOM-SAMPLED-SPLITS-GB1: Random sampled splits fitness prediction, GB1 split Configuration: ESM-1b (per AA)Task: FLIP RANDOM-SAMPLED-SPLITS-GB1: Random sampled splits fitness prediction, GB1 splitDataset subset: FLIP Random sampled splits, GB1 split (FLIP split) Trained on the split's training set with ten percent held out for validation, scored on its test set. The authors report these separately and call them optimistic: a random split of a mutational landscape leaks close relatives between train and test. Author-reported evaluation · Evaluation metadata: source checked | ||
| 0.92 spearman Unit: correlation · Direction: higher | Uncertainty: Not reported Scored: Not reported · Eligible: Not reported | source checkedflip primary benchmark evidence · Table 7, row(ESM-1b (per AA)), column(GB1) Source checking is not independent reproduction. |
| ESM-1v (per AA) on FLIP RANDOM-SAMPLED-SPLITS-GB1: Random sampled splits fitness prediction, GB1 split Configuration: ESM-1v (per AA)Task: FLIP RANDOM-SAMPLED-SPLITS-GB1: Random sampled splits fitness prediction, GB1 splitDataset subset: FLIP Random sampled splits, GB1 split (FLIP split) Trained on the split's training set with ten percent held out for validation, scored on its test set. The authors report these separately and call them optimistic: a random split of a mutational landscape leaks close relatives between train and test. Author-reported evaluation · Evaluation metadata: source checked | ||
| 0.92 spearman Unit: correlation · Direction: higher | Uncertainty: Not reported Scored: Not reported · Eligible: Not reported | source checkedflip primary benchmark evidence · Table 7, row(ESM-1v (per AA)), column(GB1) Source checking is not independent reproduction. |
| ESM-untrained (per AA) on FLIP RANDOM-SAMPLED-SPLITS-GB1: Random sampled splits fitness prediction, GB1 split Configuration: ESM-untrained (per AA)Task: FLIP RANDOM-SAMPLED-SPLITS-GB1: Random sampled splits fitness prediction, GB1 splitDataset subset: FLIP Random sampled splits, GB1 split (FLIP split) Trained on the split's training set with ten percent held out for validation, scored on its test set. The authors report these separately and call them optimistic: a random split of a mutational landscape leaks close relatives between train and test. Author-reported evaluation · Evaluation metadata: source checked | ||
| 0.79 spearman Unit: correlation · Direction: higher | Uncertainty: Not reported Scored: Not reported · Eligible: Not reported | source checkedflip primary benchmark evidence · Table 7, row(ESM-untrained (per AA)), column(GB1) Source checking is not independent reproduction. |
| Ridge on FLIP RANDOM-SAMPLED-SPLITS-GB1: Random sampled splits fitness prediction, GB1 split Method: RidgeTask: FLIP RANDOM-SAMPLED-SPLITS-GB1: Random sampled splits fitness prediction, GB1 splitDataset subset: FLIP Random sampled splits, GB1 split (FLIP split) Trained on the split's training set with ten percent held out for validation, scored on its test set. The authors report these separately and call them optimistic: a random split of a mutational landscape leaks close relatives between train and test. Author-reported evaluation · Evaluation metadata: source checked | ||
| 0.82 spearman Unit: correlation · Direction: higher | Uncertainty: Not reported Scored: Not reported · Eligible: Not reported | source checkedflip primary benchmark evidence · Table 7, row(Ridge), column(GB1) 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-flip Individual claims | flip primary benchmark evidence Table 7, column(GB1) Version: 2021 manuscript | source checked automated source review · 2026-09-18 Audit detailsPrimary-source transcription with no human sign-off and no independent reproduction. Field: Claim: flip-association-random-sampled-splits-gb1 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
- flip primary benchmark evidence · Original source · 2021 manuscript
Technical metadata and extraction receipts
Stable ID: flip-task-random-sampled-splits-gb1
- areas
- proteins-complexes
- tasks
- Random sampled splits fitness prediction, GB1 split
- metric
- Spearman correlation
- metric direction
- higher
- dataset
- FLIP Random sampled splits, GB1 split
- protocol
- Trained on the split's training set with ten percent held out for validation, scored on its test set. The authors report these separately and call them optimistic: a random split of a mutational landscape leaks close relatives between train and test.
- source locator
- Table 7, column(GB1)
- comparison panels
- id: flip-panel-random-sampled-splits-gb1; title: FLIP RANDOM-SAMPLED-SPLITS-GB1: Random sampled splits fitness prediction, GB1 split; protocol id: flip-task-random-sampled-splits-gb1; dataset id: flip-dataset-flip-random-sampled-splits-gb1-split; metric: spearman; unit: correlation; direction: higher; result ids: flip-result-esm-1b-per-aa-random-sampled-splits-gb1-spearman; flip-result-esm-1v-per-aa-random-sampled-splits-gb1-spearman; flip-result-esm-untrained-per-aa-random-sampled-splits-gb1-spearman; flip-result-ridge-random-sampled-splits-gb1-spearman; flip-result-cnn-random-sampled-splits-gb1-spearman; source ids: expansion-p3-flip; source locator: Table 7, column(GB1); context: Every method FLIP reports on Random sampled splits fitness prediction, GB1 split, scored with Spearman correlation on FLIP Random sampled splits, GB1 split.; caveats: Author-reported numbers, source checked but not independently reproduced.; The split is the task here. Figures from different splits of the same landscape are not comparable.; The random sampled splits in Table 7 are the authors' own illustration of an optimistic evaluation, not a headline result.; NA marks a baseline that cannot be applied to that landscape, and is recorded as no result rather than as a zero.; review: method: automated_source_review; date: 2026-09-18
Related records
- part of: FLIP
- subject: FLIP RANDOM-SAMPLED-SPLITS-GB1: part of discovery-benchmark-flip
- benchmark: CNN on FLIP RANDOM-SAMPLED-SPLITS-GB1: Random sampled splits fitness prediction, GB1 split
- benchmark: ESM-1b (per AA) on FLIP RANDOM-SAMPLED-SPLITS-GB1: Random sampled splits fitness prediction, GB1 split
- benchmark: ESM-1v (per AA) on FLIP RANDOM-SAMPLED-SPLITS-GB1: Random sampled splits fitness prediction, GB1 split
- benchmark: ESM-untrained (per AA) on FLIP RANDOM-SAMPLED-SPLITS-GB1: Random sampled splits fitness prediction, GB1 split
- benchmark: Ridge on FLIP RANDOM-SAMPLED-SPLITS-GB1: Random sampled splits fitness prediction, GB1 split