rewirebio.iobenchmarks
Dataset

BacBench phenotypic trait dataset

Categorical phenotypic trait labels for 24,462 genomes over 138 traits, collated from two trait catalogues with source-prefixed labels and an agreement filter, with genome sequences from NCBI GenBank.

Evaluation results

24 evaluations · 24 results. Different protocols are not a single leaderboard.

Filter evaluations

Applied filters: All linked evaluations

Exact evaluated configurations and original reported results
Tested configurationProtocol and datasetFindingEvidence and details
Configuration: Bacformer (frozen embeddings with a trained linear head)Protocol: BacBench phenotypic trait prediction, linear probe AUPRC
Dataset: BacBench phenotypic trait dataset
47.60 ± 36.13 AUPRC (linear head per trait on frozen genome embeddings, macro-averaged across classes of each trait, genus-level split, mean over traits and seeds)
unitless · higher

Uncertainty: SD 36.13

Coverage: Not reported scored / Not reported eligible

Author-reported evaluation · Source checked
Methods, coverage and source

Bacformer on BacBench phenotypic trait prediction, linear probe AUPRC

One linear model per trait on mean-pooled whole-genome embeddings: a single layer preceded by dropout of 0.1 and layer normalisation, cross-entropy loss, at most 100 epochs, early stopping patience 10 on validation AUROC, a learning-rate sweep per model across seeds and traits, and no class weighting. Scores are macro-averaged across the classes of each trait.

Aggregation: Not reported

Bacformer preprint, bioRxiv 2025.07.20.665723 version 3 · Table S16, row 'Bacformer', column 'AUPRC'
Configuration: Bacformer (frozen embeddings with a trained linear head)Protocol: BacBench phenotypic trait prediction, linear probe AUROC
Dataset: BacBench phenotypic trait dataset
71.35 ± 13.01 AUROC (linear head per trait on frozen genome embeddings, macro-averaged across classes of each trait, genus-level split, mean over traits and seeds)
unitless · higher

Uncertainty: SD 13.01

Coverage: Not reported scored / Not reported eligible

Author-reported evaluation · Source checked
Methods, coverage and source

Bacformer on BacBench phenotypic trait prediction, linear probe AUROC

One linear model per trait on mean-pooled whole-genome embeddings: a single layer preceded by dropout of 0.1 and layer normalisation, cross-entropy loss, at most 100 epochs, early stopping patience 10 on validation AUROC, a learning-rate sweep per model across seeds and traits, and no class weighting. Scores are macro-averaged across the classes of each trait.

Aggregation: Not reported

Bacformer preprint, bioRxiv 2025.07.20.665723 version 3 · Table S16, row 'Bacformer', column 'AUROC'
Configuration: Bacformer Large (frozen embeddings with a trained linear head)Protocol: BacBench phenotypic trait prediction, linear probe AUPRC
Dataset: BacBench phenotypic trait dataset
48.34 ± 36.39 AUPRC (linear head per trait on frozen genome embeddings, macro-averaged across classes of each trait, genus-level split, mean over traits and seeds)
unitless · higher

Uncertainty: SD 36.39

Coverage: Not reported scored / Not reported eligible

Author-reported evaluation · Source checked
Methods, coverage and source

Bacformer Large on BacBench phenotypic trait prediction, linear probe AUPRC

One linear model per trait on mean-pooled whole-genome embeddings: a single layer preceded by dropout of 0.1 and layer normalisation, cross-entropy loss, at most 100 epochs, early stopping patience 10 on validation AUROC, a learning-rate sweep per model across seeds and traits, and no class weighting. Scores are macro-averaged across the classes of each trait.

Aggregation: Not reported

Bacformer preprint, bioRxiv 2025.07.20.665723 version 3 · Table S16, row 'Bacformer Large', column 'AUPRC'
Configuration: Bacformer Large (frozen embeddings with a trained linear head)Protocol: BacBench phenotypic trait prediction, linear probe AUROC
Dataset: BacBench phenotypic trait dataset
73.33 ± 12.13 AUROC (linear head per trait on frozen genome embeddings, macro-averaged across classes of each trait, genus-level split, mean over traits and seeds)
unitless · higher

Uncertainty: SD 12.13

Coverage: Not reported scored / Not reported eligible

Author-reported evaluation · Source checked
Methods, coverage and source

Bacformer Large on BacBench phenotypic trait prediction, linear probe AUROC

One linear model per trait on mean-pooled whole-genome embeddings: a single layer preceded by dropout of 0.1 and layer normalisation, cross-entropy loss, at most 100 epochs, early stopping patience 10 on validation AUROC, a learning-rate sweep per model across seeds and traits, and no class weighting. Scores are macro-averaged across the classes of each trait.

Aggregation: Not reported

Bacformer preprint, bioRxiv 2025.07.20.665723 version 3 · Table S16, row 'Bacformer Large', column 'AUROC'
Configuration: DNABERT-2 (frozen embeddings with a trained linear head)Protocol: BacBench phenotypic trait prediction, linear probe AUPRC
Dataset: BacBench phenotypic trait dataset
41.93 ± 34.31 AUPRC (linear head per trait on frozen genome embeddings, macro-averaged across classes of each trait, genus-level split, mean over traits and seeds)
unitless · higher

Uncertainty: SD 34.31

Coverage: Not reported scored / Not reported eligible

Author-reported evaluation · Source checked
Methods, coverage and source

DNABERT-2 on BacBench phenotypic trait prediction, linear probe AUPRC

One linear model per trait on mean-pooled whole-genome embeddings: a single layer preceded by dropout of 0.1 and layer normalisation, cross-entropy loss, at most 100 epochs, early stopping patience 10 on validation AUROC, a learning-rate sweep per model across seeds and traits, and no class weighting. Scores are macro-averaged across the classes of each trait.

Aggregation: Not reported

Bacformer preprint, bioRxiv 2025.07.20.665723 version 3 · Table S16, row 'DNABERT-2', column 'AUPRC'
Configuration: DNABERT-2 (frozen embeddings with a trained linear head)Protocol: BacBench phenotypic trait prediction, linear probe AUROC
Dataset: BacBench phenotypic trait dataset
63.81 ± 11.15 AUROC (linear head per trait on frozen genome embeddings, macro-averaged across classes of each trait, genus-level split, mean over traits and seeds)
unitless · higher

Uncertainty: SD 11.15

Coverage: Not reported scored / Not reported eligible

Author-reported evaluation · Source checked
Methods, coverage and source

DNABERT-2 on BacBench phenotypic trait prediction, linear probe AUROC

One linear model per trait on mean-pooled whole-genome embeddings: a single layer preceded by dropout of 0.1 and layer normalisation, cross-entropy loss, at most 100 epochs, early stopping patience 10 on validation AUROC, a learning-rate sweep per model across seeds and traits, and no class weighting. Scores are macro-averaged across the classes of each trait.

Aggregation: Not reported

Bacformer preprint, bioRxiv 2025.07.20.665723 version 3 · Table S16, row 'DNABERT-2', column 'AUROC'
Configuration: ESM-2 (frozen embeddings with a trained linear head)Protocol: BacBench phenotypic trait prediction, linear probe AUPRC
Dataset: BacBench phenotypic trait dataset
45.42 ± 35.54 AUPRC (linear head per trait on frozen genome embeddings, macro-averaged across classes of each trait, genus-level split, mean over traits and seeds)
unitless · higher

Uncertainty: SD 35.54

Coverage: Not reported scored / Not reported eligible

Author-reported evaluation · Source checked
Methods, coverage and source

ESM-2 on BacBench phenotypic trait prediction, linear probe AUPRC

One linear model per trait on mean-pooled whole-genome embeddings: a single layer preceded by dropout of 0.1 and layer normalisation, cross-entropy loss, at most 100 epochs, early stopping patience 10 on validation AUROC, a learning-rate sweep per model across seeds and traits, and no class weighting. Scores are macro-averaged across the classes of each trait.

Aggregation: Not reported

Bacformer preprint, bioRxiv 2025.07.20.665723 version 3 · Table S16, row 'ESM-2', column 'AUPRC'
Configuration: ESM-2 (frozen embeddings with a trained linear head)Protocol: BacBench phenotypic trait prediction, linear probe AUROC
Dataset: BacBench phenotypic trait dataset
68.68 ± 12.46 AUROC (linear head per trait on frozen genome embeddings, macro-averaged across classes of each trait, genus-level split, mean over traits and seeds)
unitless · higher

Uncertainty: SD 12.46

Coverage: Not reported scored / Not reported eligible

Author-reported evaluation · Source checked
Methods, coverage and source

ESM-2 on BacBench phenotypic trait prediction, linear probe AUROC

One linear model per trait on mean-pooled whole-genome embeddings: a single layer preceded by dropout of 0.1 and layer normalisation, cross-entropy loss, at most 100 epochs, early stopping patience 10 on validation AUROC, a learning-rate sweep per model across seeds and traits, and no class weighting. Scores are macro-averaged across the classes of each trait.

Aggregation: Not reported

Bacformer preprint, bioRxiv 2025.07.20.665723 version 3 · Table S16, row 'ESM-2', column 'AUROC'
Configuration: ESM-C (frozen embeddings with a trained linear head)Protocol: BacBench phenotypic trait prediction, linear probe AUPRC
Dataset: BacBench phenotypic trait dataset
45.36 ± 35.47 AUPRC (linear head per trait on frozen genome embeddings, macro-averaged across classes of each trait, genus-level split, mean over traits and seeds)
unitless · higher

Uncertainty: SD 35.47

Coverage: Not reported scored / Not reported eligible

Author-reported evaluation · Source checked
Methods, coverage and source

ESM-C on BacBench phenotypic trait prediction, linear probe AUPRC

One linear model per trait on mean-pooled whole-genome embeddings: a single layer preceded by dropout of 0.1 and layer normalisation, cross-entropy loss, at most 100 epochs, early stopping patience 10 on validation AUROC, a learning-rate sweep per model across seeds and traits, and no class weighting. Scores are macro-averaged across the classes of each trait.

Aggregation: Not reported

Bacformer preprint, bioRxiv 2025.07.20.665723 version 3 · Table S16, row 'ESM-C', column 'AUPRC'
Configuration: ESM-C (frozen embeddings with a trained linear head)Protocol: BacBench phenotypic trait prediction, linear probe AUROC
Dataset: BacBench phenotypic trait dataset
69.06 ± 12.07 AUROC (linear head per trait on frozen genome embeddings, macro-averaged across classes of each trait, genus-level split, mean over traits and seeds)
unitless · higher

Uncertainty: SD 12.07

Coverage: Not reported scored / Not reported eligible

Author-reported evaluation · Source checked
Methods, coverage and source

ESM-C on BacBench phenotypic trait prediction, linear probe AUROC

One linear model per trait on mean-pooled whole-genome embeddings: a single layer preceded by dropout of 0.1 and layer normalisation, cross-entropy loss, at most 100 epochs, early stopping patience 10 on validation AUROC, a learning-rate sweep per model across seeds and traits, and no class weighting. Scores are macro-averaged across the classes of each trait.

Aggregation: Not reported

Bacformer preprint, bioRxiv 2025.07.20.665723 version 3 · Table S16, row 'ESM-C', column 'AUROC'
Configuration: Evo-2 (frozen embeddings with a trained linear head)Protocol: BacBench phenotypic trait prediction, linear probe AUPRC
Dataset: BacBench phenotypic trait dataset
24.96 ± 20.52 AUPRC (linear head per trait on frozen genome embeddings, macro-averaged across classes of each trait, genus-level split, mean over traits and seeds)
unitless · higher

Uncertainty: SD 20.52

Coverage: Not reported scored / Not reported eligible

Author-reported evaluation · Source checked
Methods, coverage and source

Evo-2 on BacBench phenotypic trait prediction, linear probe AUPRC

One linear model per trait on mean-pooled whole-genome embeddings: a single layer preceded by dropout of 0.1 and layer normalisation, cross-entropy loss, at most 100 epochs, early stopping patience 10 on validation AUROC, a learning-rate sweep per model across seeds and traits, and no class weighting. Scores are macro-averaged across the classes of each trait.

Aggregation: Not reported

Bacformer preprint, bioRxiv 2025.07.20.665723 version 3 · Table S16, row 'Evo-2', column 'AUPRC'
Configuration: Evo-2 (frozen embeddings with a trained linear head)Protocol: BacBench phenotypic trait prediction, linear probe AUROC
Dataset: BacBench phenotypic trait dataset
55.41 ± 8.85 AUROC (linear head per trait on frozen genome embeddings, macro-averaged across classes of each trait, genus-level split, mean over traits and seeds)
unitless · higher

Uncertainty: SD 8.85

Coverage: Not reported scored / Not reported eligible

Author-reported evaluation · Source checked
Methods, coverage and source

Evo-2 on BacBench phenotypic trait prediction, linear probe AUROC

One linear model per trait on mean-pooled whole-genome embeddings: a single layer preceded by dropout of 0.1 and layer normalisation, cross-entropy loss, at most 100 epochs, early stopping patience 10 on validation AUROC, a learning-rate sweep per model across seeds and traits, and no class weighting. Scores are macro-averaged across the classes of each trait.

Aggregation: Not reported

Bacformer preprint, bioRxiv 2025.07.20.665723 version 3 · Table S16, row 'Evo-2', column 'AUROC'
Configuration: Evo (frozen embeddings with a trained linear head)Protocol: BacBench phenotypic trait prediction, linear probe AUPRC
Dataset: BacBench phenotypic trait dataset
39.55 ± 33.36 AUPRC (linear head per trait on frozen genome embeddings, macro-averaged across classes of each trait, genus-level split, mean over traits and seeds)
unitless · higher

Uncertainty: SD 33.36

Coverage: Not reported scored / Not reported eligible

Author-reported evaluation · Source checked
Methods, coverage and source

Evo on BacBench phenotypic trait prediction, linear probe AUPRC

One linear model per trait on mean-pooled whole-genome embeddings: a single layer preceded by dropout of 0.1 and layer normalisation, cross-entropy loss, at most 100 epochs, early stopping patience 10 on validation AUROC, a learning-rate sweep per model across seeds and traits, and no class weighting. Scores are macro-averaged across the classes of each trait.

Aggregation: Not reported

Bacformer preprint, bioRxiv 2025.07.20.665723 version 3 · Table S16, row 'Evo', column 'AUPRC'
Configuration: Evo (frozen embeddings with a trained linear head)Protocol: BacBench phenotypic trait prediction, linear probe AUROC
Dataset: BacBench phenotypic trait dataset
60.19 ± 11.04 AUROC (linear head per trait on frozen genome embeddings, macro-averaged across classes of each trait, genus-level split, mean over traits and seeds)
unitless · higher

Uncertainty: SD 11.04

Coverage: Not reported scored / Not reported eligible

Author-reported evaluation · Source checked
Methods, coverage and source

Evo on BacBench phenotypic trait prediction, linear probe AUROC

One linear model per trait on mean-pooled whole-genome embeddings: a single layer preceded by dropout of 0.1 and layer normalisation, cross-entropy loss, at most 100 epochs, early stopping patience 10 on validation AUROC, a learning-rate sweep per model across seeds and traits, and no class weighting. Scores are macro-averaged across the classes of each trait.

Aggregation: Not reported

Bacformer preprint, bioRxiv 2025.07.20.665723 version 3 · Table S16, row 'Evo', column 'AUROC'
Configuration: gLM2 (frozen embeddings with a trained linear head)Protocol: BacBench phenotypic trait prediction, linear probe AUPRC
Dataset: BacBench phenotypic trait dataset
43.77 ± 34.45 AUPRC (linear head per trait on frozen genome embeddings, macro-averaged across classes of each trait, genus-level split, mean over traits and seeds)
unitless · higher

Uncertainty: SD 34.45

Coverage: Not reported scored / Not reported eligible

Author-reported evaluation · Source checked
Methods, coverage and source

gLM2 on BacBench phenotypic trait prediction, linear probe AUPRC

One linear model per trait on mean-pooled whole-genome embeddings: a single layer preceded by dropout of 0.1 and layer normalisation, cross-entropy loss, at most 100 epochs, early stopping patience 10 on validation AUROC, a learning-rate sweep per model across seeds and traits, and no class weighting. Scores are macro-averaged across the classes of each trait.

Aggregation: Not reported

Bacformer preprint, bioRxiv 2025.07.20.665723 version 3 · Table S16, row 'gLM2', column 'AUPRC'
Configuration: gLM2 (frozen embeddings with a trained linear head)Protocol: BacBench phenotypic trait prediction, linear probe AUROC
Dataset: BacBench phenotypic trait dataset
65.74 ± 11.47 AUROC (linear head per trait on frozen genome embeddings, macro-averaged across classes of each trait, genus-level split, mean over traits and seeds)
unitless · higher

Uncertainty: SD 11.47

Coverage: Not reported scored / Not reported eligible

Author-reported evaluation · Source checked
Methods, coverage and source

gLM2 on BacBench phenotypic trait prediction, linear probe AUROC

One linear model per trait on mean-pooled whole-genome embeddings: a single layer preceded by dropout of 0.1 and layer normalisation, cross-entropy loss, at most 100 epochs, early stopping patience 10 on validation AUROC, a learning-rate sweep per model across seeds and traits, and no class weighting. Scores are macro-averaged across the classes of each trait.

Aggregation: Not reported

Bacformer preprint, bioRxiv 2025.07.20.665723 version 3 · Table S16, row 'gLM2', column 'AUROC'
Configuration: Mistral-DNA (frozen embeddings with a trained linear head)Protocol: BacBench phenotypic trait prediction, linear probe AUPRC
Dataset: BacBench phenotypic trait dataset
39.56 ± 33.16 AUPRC (linear head per trait on frozen genome embeddings, macro-averaged across classes of each trait, genus-level split, mean over traits and seeds)
unitless · higher

Uncertainty: SD 33.16

Coverage: Not reported scored / Not reported eligible

Author-reported evaluation · Source checked
Methods, coverage and source

Mistral-DNA on BacBench phenotypic trait prediction, linear probe AUPRC

One linear model per trait on mean-pooled whole-genome embeddings: a single layer preceded by dropout of 0.1 and layer normalisation, cross-entropy loss, at most 100 epochs, early stopping patience 10 on validation AUROC, a learning-rate sweep per model across seeds and traits, and no class weighting. Scores are macro-averaged across the classes of each trait.

Aggregation: Not reported

Bacformer preprint, bioRxiv 2025.07.20.665723 version 3 · Table S16, row 'Mistral-DNA', column 'AUPRC'
Configuration: Mistral-DNA (frozen embeddings with a trained linear head)Protocol: BacBench phenotypic trait prediction, linear probe AUROC
Dataset: BacBench phenotypic trait dataset
60.33 ± 9.88 AUROC (linear head per trait on frozen genome embeddings, macro-averaged across classes of each trait, genus-level split, mean over traits and seeds)
unitless · higher

Uncertainty: SD 9.88

Coverage: Not reported scored / Not reported eligible

Author-reported evaluation · Source checked
Methods, coverage and source

Mistral-DNA on BacBench phenotypic trait prediction, linear probe AUROC

One linear model per trait on mean-pooled whole-genome embeddings: a single layer preceded by dropout of 0.1 and layer normalisation, cross-entropy loss, at most 100 epochs, early stopping patience 10 on validation AUROC, a learning-rate sweep per model across seeds and traits, and no class weighting. Scores are macro-averaged across the classes of each trait.

Aggregation: Not reported

Bacformer preprint, bioRxiv 2025.07.20.665723 version 3 · Table S16, row 'Mistral-DNA', column 'AUROC'
Configuration: Nucleotide Transformer (frozen embeddings with a trained linear head)Protocol: BacBench phenotypic trait prediction, linear probe AUPRC
Dataset: BacBench phenotypic trait dataset
43.05 ± 34.91 AUPRC (linear head per trait on frozen genome embeddings, macro-averaged across classes of each trait, genus-level split, mean over traits and seeds)
unitless · higher

Uncertainty: SD 34.91

Coverage: Not reported scored / Not reported eligible

Author-reported evaluation · Source checked
Methods, coverage and source

Nucleotide Transformer on BacBench phenotypic trait prediction, linear probe AUPRC

One linear model per trait on mean-pooled whole-genome embeddings: a single layer preceded by dropout of 0.1 and layer normalisation, cross-entropy loss, at most 100 epochs, early stopping patience 10 on validation AUROC, a learning-rate sweep per model across seeds and traits, and no class weighting. Scores are macro-averaged across the classes of each trait.

Aggregation: Not reported

Bacformer preprint, bioRxiv 2025.07.20.665723 version 3 · Table S16, row 'Nucleotide Transformer', column 'AUPRC'
Configuration: Nucleotide Transformer (frozen embeddings with a trained linear head)Protocol: BacBench phenotypic trait prediction, linear probe AUROC
Dataset: BacBench phenotypic trait dataset
65.98 ± 11.54 AUROC (linear head per trait on frozen genome embeddings, macro-averaged across classes of each trait, genus-level split, mean over traits and seeds)
unitless · higher

Uncertainty: SD 11.54

Coverage: Not reported scored / Not reported eligible

Author-reported evaluation · Source checked
Methods, coverage and source

Nucleotide Transformer on BacBench phenotypic trait prediction, linear probe AUROC

One linear model per trait on mean-pooled whole-genome embeddings: a single layer preceded by dropout of 0.1 and layer normalisation, cross-entropy loss, at most 100 epochs, early stopping patience 10 on validation AUROC, a learning-rate sweep per model across seeds and traits, and no class weighting. Scores are macro-averaged across the classes of each trait.

Aggregation: Not reported

Bacformer preprint, bioRxiv 2025.07.20.665723 version 3 · Table S16, row 'Nucleotide Transformer', column 'AUROC'
Configuration: ProkBERT (frozen embeddings with a trained linear head)Protocol: BacBench phenotypic trait prediction, linear probe AUPRC
Dataset: BacBench phenotypic trait dataset
44.58 ± 34.78 AUPRC (linear head per trait on frozen genome embeddings, macro-averaged across classes of each trait, genus-level split, mean over traits and seeds)
unitless · higher

Uncertainty: SD 34.78

Coverage: Not reported scored / Not reported eligible

Author-reported evaluation · Source checked
Methods, coverage and source

ProkBERT on BacBench phenotypic trait prediction, linear probe AUPRC

One linear model per trait on mean-pooled whole-genome embeddings: a single layer preceded by dropout of 0.1 and layer normalisation, cross-entropy loss, at most 100 epochs, early stopping patience 10 on validation AUROC, a learning-rate sweep per model across seeds and traits, and no class weighting. Scores are macro-averaged across the classes of each trait.

Aggregation: Not reported

Bacformer preprint, bioRxiv 2025.07.20.665723 version 3 · Table S16, row 'ProkBERT', column 'AUPRC'
Configuration: ProkBERT (frozen embeddings with a trained linear head)Protocol: BacBench phenotypic trait prediction, linear probe AUROC
Dataset: BacBench phenotypic trait dataset
67.58 ± 11.94 AUROC (linear head per trait on frozen genome embeddings, macro-averaged across classes of each trait, genus-level split, mean over traits and seeds)
unitless · higher

Uncertainty: SD 11.94

Coverage: Not reported scored / Not reported eligible

Author-reported evaluation · Source checked
Methods, coverage and source

ProkBERT on BacBench phenotypic trait prediction, linear probe AUROC

One linear model per trait on mean-pooled whole-genome embeddings: a single layer preceded by dropout of 0.1 and layer normalisation, cross-entropy loss, at most 100 epochs, early stopping patience 10 on validation AUROC, a learning-rate sweep per model across seeds and traits, and no class weighting. Scores are macro-averaged across the classes of each trait.

Aggregation: Not reported

Bacformer preprint, bioRxiv 2025.07.20.665723 version 3 · Table S16, row 'ProkBERT', column 'AUROC'
Configuration: ProtBERT (frozen embeddings with a trained linear head)Protocol: BacBench phenotypic trait prediction, linear probe AUPRC
Dataset: BacBench phenotypic trait dataset
45.24 ± 34.98 AUPRC (linear head per trait on frozen genome embeddings, macro-averaged across classes of each trait, genus-level split, mean over traits and seeds)
unitless · higher

Uncertainty: SD 34.98

Coverage: Not reported scored / Not reported eligible

Author-reported evaluation · Source checked
Methods, coverage and source

ProtBERT on BacBench phenotypic trait prediction, linear probe AUPRC

One linear model per trait on mean-pooled whole-genome embeddings: a single layer preceded by dropout of 0.1 and layer normalisation, cross-entropy loss, at most 100 epochs, early stopping patience 10 on validation AUROC, a learning-rate sweep per model across seeds and traits, and no class weighting. Scores are macro-averaged across the classes of each trait.

Aggregation: Not reported

Bacformer preprint, bioRxiv 2025.07.20.665723 version 3 · Table S16, row 'ProtBERT', column 'AUPRC'
Configuration: ProtBERT (frozen embeddings with a trained linear head)Protocol: BacBench phenotypic trait prediction, linear probe AUROC
Dataset: BacBench phenotypic trait dataset
68.40 ± 12.05 AUROC (linear head per trait on frozen genome embeddings, macro-averaged across classes of each trait, genus-level split, mean over traits and seeds)
unitless · higher

Uncertainty: SD 12.05

Coverage: Not reported scored / Not reported eligible

Author-reported evaluation · Source checked
Methods, coverage and source

ProtBERT on BacBench phenotypic trait prediction, linear probe AUROC

One linear model per trait on mean-pooled whole-genome embeddings: a single layer preceded by dropout of 0.1 and layer normalisation, cross-entropy loss, at most 100 epochs, early stopping patience 10 on validation AUROC, a learning-rate sweep per model across seeds and traits, and no class weighting. Scores are macro-averaged across the classes of each trait.

Aggregation: Not reported

Bacformer preprint, bioRxiv 2025.07.20.665723 version 3 · Table S16, row 'ProtBERT', column 'AUROC'

Source checking is not independent reproduction. Release 2026-11-21-adc15377fab2.

Research readiness

0 of 4 readiness checks met. These checks assess whether the evidence supports a reproducible investigation; a source-checked score alone does not meet them.

Readiness checks, gaps and artifacts

Release 2026-11-21-adc15377fab2 · Evidence verified: Not verified

Evidence incomplete

Replay metrics

Exact outcomes, predictions, identifiers and evaluator are connected.

Missing or unresolved evidence

  • No verified artifact manifest is linked to this exact record.
  • File checksums match the recorded files: not yet verified
  • Predictions are matched to the right samples: not yet verified
  • Score meaning and direction are confirmed: not yet verified
  • Metrics are recomputed from the saved predictions: not yet verified

Verified: Not verified

Evidence incomplete

Investigate discrepancies

Replay evidence includes annotations and an assessment of dependence. Unknown independence permits descriptive analysis only.

Missing or unresolved evidence

  • No verified artifact manifest is linked to this exact record.
  • File checksums match the recorded files: not yet verified
  • Predictions are matched to the right samples: not yet verified
  • Score meaning and direction are confirmed: not yet verified
  • Metrics are recomputed from the saved predictions: not yet verified
  • Sample annotations are recorded: not yet verified
  • Dependence between samples is assessed: not yet verified

Verified: Not verified

Evidence incomplete

Run locally

A pinned recipe describes the inputs, environment and resource requirements.

Missing or unresolved evidence

  • No verified artifact manifest is linked to this exact record.
  • File checksums match the recorded files: not yet verified
  • Predictions are matched to the right samples: not yet verified
  • Score meaning and direction are confirmed: not yet verified
  • A pinned run recipe exists: not yet verified
  • Compute requirements are estimated: not yet verified

Verified: Not verified

Evidence incomplete

Validate independently

Separate data and exposure records support an independent test.

Missing or unresolved evidence

  • No verified artifact manifest is linked to this exact record.
  • File checksums match the recorded files: not yet verified
  • Predictions are matched to the right samples: not yet verified
  • Score meaning and direction are confirmed: not yet verified
  • Independent validation data exist: not yet verified
  • Overlap with training data is checked: not yet verified

Verified: Not verified

Readiness describes the evidence in this release. Availability on your computer is checked separately when an investigation runs. Existing data exposure can prevent independent validation even when files are available.

Artifacts and reproduction

No verified artifact manifest is connected to this record yet. The gaps above identify what is needed before analysis can begin.

Dataset and evaluation context

A dataset supplies biological observations. The evaluation protocol defines how those observations are split, used and scored.

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.

24 evidence rows matching the loaded filters

Claims, original sources and review scope · Release 2026-11-21-adc15377fab2
Property and statementOriginal source and locationReview and provenance
Access
Published as Hugging Face datasets in protein-sequence and DNA modalities.
Context-only references
BacBench repository README at commit 64b7d45

Original source ↗

Methods 'Phenotypic traits prediction'; BacBench README 'Datasets'

Shared locator for this statement’s cited sources; not a separate locator for each citation.

Version: Commit 64b7d45823b2ff2ef88458e3aa235caa461a12aa
Retrieved: 2026-11-15T23:09:36Z

not individually reviewed

No individual claim review recorded

Audit details

Field: attributes.access

Source artifact SHA-256: c9d28efef5b4cc30adbcc9de224c350100f5aa2a15a025beaf10dcdeb86cd93e

Hash scope: Hash scope not separately documented; inspect source record

Inspected artifact

Access
Published as Hugging Face datasets in protein-sequence and DNA modalities.
Context-only references
Bacformer preprint, bioRxiv 2025.07.20.665723 version 3

Original source ↗

Methods 'Phenotypic traits prediction'; BacBench README 'Datasets'

Shared locator for this statement’s cited sources; not a separate locator for each citation.

Version: Version 3, posted 17 August 2026 according to the page banner; the bioRxiv API lists version 3 with date 2026-08-19
Retrieved: 2026-11-15T22:55:00Z

not individually reviewed

No individual claim review recorded

Audit details

Field: attributes.access

Source artifact SHA-256: d86de3315063e038ec8d5bed60b7b1a916c137ee60973e55d9d5c82f078d8dfd

Hash scope: The hash is of one download. bioRxiv regenerates this PDF on request, so the container bytes differ between downloads: a second download minutes later gave 08fcc83af256845d942345fc6b63c64911878733acd028edab5e9307d63319db. The pdftotext -layout text layer of both downloads is byte-identical (SHA-256 86636e89b167d7304b3a83ee7f159cad0a9bdbe97019588ce836087b8fe3b0b2), and that text layer, archived gzipped in artifacts/ with SHA-256 0afdefd78a410d036d1c4f17e33bdf481e8d76241ee02f4c84c7c4be91ee72aa, is what every extracted value was read from.

Assay
Trait annotations collated from two published catalogues; traits present in both were kept only when Cohen's kappa between the sources reached 0.5.
Context-only references
BacBench repository README at commit 64b7d45

Original source ↗

Methods 'Phenotypic traits prediction'; BacBench README 'Datasets'

Shared locator for this statement’s cited sources; not a separate locator for each citation.

Version: Commit 64b7d45823b2ff2ef88458e3aa235caa461a12aa
Retrieved: 2026-11-15T23:09:36Z

not individually reviewed

No individual claim review recorded

Audit details

Field: attributes.assay

Source artifact SHA-256: c9d28efef5b4cc30adbcc9de224c350100f5aa2a15a025beaf10dcdeb86cd93e

Hash scope: Hash scope not separately documented; inspect source record

Inspected artifact

Assay
Trait annotations collated from two published catalogues; traits present in both were kept only when Cohen's kappa between the sources reached 0.5.
Context-only references
Bacformer preprint, bioRxiv 2025.07.20.665723 version 3

Original source ↗

Methods 'Phenotypic traits prediction'; BacBench README 'Datasets'

Shared locator for this statement’s cited sources; not a separate locator for each citation.

Version: Version 3, posted 17 August 2026 according to the page banner; the bioRxiv API lists version 3 with date 2026-08-19
Retrieved: 2026-11-15T22:55:00Z

not individually reviewed

No individual claim review recorded

Audit details

Field: attributes.assay

Source artifact SHA-256: d86de3315063e038ec8d5bed60b7b1a916c137ee60973e55d9d5c82f078d8dfd

Hash scope: The hash is of one download. bioRxiv regenerates this PDF on request, so the container bytes differ between downloads: a second download minutes later gave 08fcc83af256845d942345fc6b63c64911878733acd028edab5e9307d63319db. The pdftotext -layout text layer of both downloads is byte-identical (SHA-256 86636e89b167d7304b3a83ee7f159cad0a9bdbe97019588ce836087b8fe3b0b2), and that text layer, archived gzipped in artifacts/ with SHA-256 0afdefd78a410d036d1c4f17e33bdf481e8d76241ee02f4c84c7c4be91ee72aa, is what every extracted value was read from.

Label semantics
Each label is a category of one trait, prefixed with the catalogue it came from so that collisions between the two catalogues are explicit.
Context-only references
BacBench repository README at commit 64b7d45

Original source ↗

Methods 'Phenotypic traits prediction'; BacBench README 'Datasets'

Shared locator for this statement’s cited sources; not a separate locator for each citation.

Version: Commit 64b7d45823b2ff2ef88458e3aa235caa461a12aa
Retrieved: 2026-11-15T23:09:36Z

not individually reviewed

No individual claim review recorded

Audit details

Field: attributes.label_semantics

Source artifact SHA-256: c9d28efef5b4cc30adbcc9de224c350100f5aa2a15a025beaf10dcdeb86cd93e

Hash scope: Hash scope not separately documented; inspect source record

Inspected artifact

Label semantics
Each label is a category of one trait, prefixed with the catalogue it came from so that collisions between the two catalogues are explicit.
Context-only references
Bacformer preprint, bioRxiv 2025.07.20.665723 version 3

Original source ↗

Methods 'Phenotypic traits prediction'; BacBench README 'Datasets'

Shared locator for this statement’s cited sources; not a separate locator for each citation.

Version: Version 3, posted 17 August 2026 according to the page banner; the bioRxiv API lists version 3 with date 2026-08-19
Retrieved: 2026-11-15T22:55:00Z

not individually reviewed

No individual claim review recorded

Audit details

Field: attributes.label_semantics

Source artifact SHA-256: d86de3315063e038ec8d5bed60b7b1a916c137ee60973e55d9d5c82f078d8dfd

Hash scope: The hash is of one download. bioRxiv regenerates this PDF on request, so the container bytes differ between downloads: a second download minutes later gave 08fcc83af256845d942345fc6b63c64911878733acd028edab5e9307d63319db. The pdftotext -layout text layer of both downloads is byte-identical (SHA-256 86636e89b167d7304b3a83ee7f159cad0a9bdbe97019588ce836087b8fe3b0b2), and that text layer, archived gzipped in artifacts/ with SHA-256 0afdefd78a410d036d1c4f17e33bdf481e8d76241ee02f4c84c7c4be91ee72aa, is what every extracted value was read from.

Population
138 categorical traits across 24,462 genomes, after keeping traits with at least 500 genomes and classes with at least 50 samples
Context-only references
BacBench repository README at commit 64b7d45

Original source ↗

Methods 'Phenotypic traits prediction'; BacBench README 'Datasets'

Shared locator for this statement’s cited sources; not a separate locator for each citation.

Version: Commit 64b7d45823b2ff2ef88458e3aa235caa461a12aa
Retrieved: 2026-11-15T23:09:36Z

not individually reviewed

No individual claim review recorded

Audit details

Field: attributes.population

Source artifact SHA-256: c9d28efef5b4cc30adbcc9de224c350100f5aa2a15a025beaf10dcdeb86cd93e

Hash scope: Hash scope not separately documented; inspect source record

Inspected artifact

Population
138 categorical traits across 24,462 genomes, after keeping traits with at least 500 genomes and classes with at least 50 samples
Context-only references
Bacformer preprint, bioRxiv 2025.07.20.665723 version 3

Original source ↗

Methods 'Phenotypic traits prediction'; BacBench README 'Datasets'

Shared locator for this statement’s cited sources; not a separate locator for each citation.

Version: Version 3, posted 17 August 2026 according to the page banner; the bioRxiv API lists version 3 with date 2026-08-19
Retrieved: 2026-11-15T22:55:00Z

not individually reviewed

No individual claim review recorded

Audit details

Field: attributes.population

Source artifact SHA-256: d86de3315063e038ec8d5bed60b7b1a916c137ee60973e55d9d5c82f078d8dfd

Hash scope: The hash is of one download. bioRxiv regenerates this PDF on request, so the container bytes differ between downloads: a second download minutes later gave 08fcc83af256845d942345fc6b63c64911878733acd028edab5e9307d63319db. The pdftotext -layout text layer of both downloads is byte-identical (SHA-256 86636e89b167d7304b3a83ee7f159cad0a9bdbe97019588ce836087b8fe3b0b2), and that text layer, archived gzipped in artifacts/ with SHA-256 0afdefd78a410d036d1c4f17e33bdf481e8d76241ee02f4c84c7c4be91ee72aa, is what every extracted value was read from.

Repository
https://huggingface.co/datasets/macwiatrak/bacbench-phenotypic-traits-protein-sequences
Context-only references
BacBench repository README at commit 64b7d45

Original source ↗

Methods 'Phenotypic traits prediction'; BacBench README 'Datasets'

Shared locator for this statement’s cited sources; not a separate locator for each citation.

Version: Commit 64b7d45823b2ff2ef88458e3aa235caa461a12aa
Retrieved: 2026-11-15T23:09:36Z

not individually reviewed

No individual claim review recorded

Audit details

Field: attributes.repository

Source artifact SHA-256: c9d28efef5b4cc30adbcc9de224c350100f5aa2a15a025beaf10dcdeb86cd93e

Hash scope: Hash scope not separately documented; inspect source record

Inspected artifact

Repository
https://huggingface.co/datasets/macwiatrak/bacbench-phenotypic-traits-protein-sequences
Context-only references
Bacformer preprint, bioRxiv 2025.07.20.665723 version 3

Original source ↗

Methods 'Phenotypic traits prediction'; BacBench README 'Datasets'

Shared locator for this statement’s cited sources; not a separate locator for each citation.

Version: Version 3, posted 17 August 2026 according to the page banner; the bioRxiv API lists version 3 with date 2026-08-19
Retrieved: 2026-11-15T22:55:00Z

not individually reviewed

No individual claim review recorded

Audit details

Field: attributes.repository

Source artifact SHA-256: d86de3315063e038ec8d5bed60b7b1a916c137ee60973e55d9d5c82f078d8dfd

Hash scope: The hash is of one download. bioRxiv regenerates this PDF on request, so the container bytes differ between downloads: a second download minutes later gave 08fcc83af256845d942345fc6b63c64911878733acd028edab5e9307d63319db. The pdftotext -layout text layer of both downloads is byte-identical (SHA-256 86636e89b167d7304b3a83ee7f159cad0a9bdbe97019588ce836087b8fe3b0b2), and that text layer, archived gzipped in artifacts/ with SHA-256 0afdefd78a410d036d1c4f17e33bdf481e8d76241ee02f4c84c7c4be91ee72aa, is what every extracted value was read from.

How to read the evidence

Sources and history

Release 2026-11-21-adc15377fab2 · Record review: source checked

2 source records and release historyDownload this release (gzip)
Technical metadata and record identifiers

Technical metadata for this record is tracked under stable ID bacformer-20261115-data-phenotypic-traits.

Source locator: Methods 'Phenotypic traits prediction'; BacBench README 'Datasets'

Related records

Suggest a correction