This MFASS pipeline uses a frozen DNABERT-2 encoder with an assay-specific logistic-regression head. It is distinct from the base encoder and from end-to-end fine-tuning.
Explanatory profile: source reviewed · Automated source review, 2026-09-16. This does not change the review status of its results.
| Property | Description and evidence |
|---|---|
| Encoder revision | b5ae377faa374ee160eec1c27b8494436cc94451MFASS v2 pinned rewire artifacts · benchmarks/mfass/README.md, DNABERT-2 protocol; results/dnabert2-117m-frozen-pair-logreg.json, config at bee9133b83f3aedaf2bbb9013f1875515845607e |
| Head | StandardScaler + balanced L2 logistic regression, C=0.1MFASS v2 pinned rewire artifacts · benchmarks/mfass/README.md, DNABERT-2 protocol; results/dnabert2-117m-frozen-pair-logreg.json, config at bee9133b83f3aedaf2bbb9013f1875515845607e |
| Context | 170 bases for each reference and mutant sequenceMFASS v2 pinned rewire artifacts · benchmarks/mfass/README.md, DNABERT-2 protocol; results/dnabert2-117m-frozen-pair-logreg.json, config at bee9133b83f3aedaf2bbb9013f1875515845607e |
| Model type | Not extracted or verified for this record. |
| Known versions | Not extracted or verified for this record. |
| Training data | Not extracted or verified for this record. |
| Access | Not extracted or verified for this record. |
| Code licence | Not extracted or verified for this record. |
| Weights licence | Not extracted or verified for this record. |
Schematic of the pinned MFASS configuration; this does not generalise to every member of the model family.
Validated 170-base reference and mutant sequences are embedded separately. Attention-mask means of final hidden states are combined as reference plus mutant-minus-reference. Standardisation and balanced L2 logistic regression are fitted only on the training arm.
MFASS v2 pinned rewire artifacts · benchmarks/mfass/README.md, DNABERT-2 protocol; results/dnabert2-117m-frozen-pair-logreg.json, config at bee9133b83f3aedaf2bbb9013f1875515845607eRelease 2026-09-16-d74d282221a9 · 1 evaluation · 3 metric rows. Different protocols are not a single leaderboard.
| Metric and finding | Coverage and uncertainty | Evidence |
|---|---|---|
| DNABERT-2 117M · frozen pair embeddings on MFASS v2 Model: DNABERT-2 117M · frozen pair embeddings · Benchmark: MFASS v2 · Dataset: MFASS v2 eligible assay cohort Masked mean of frozen last hidden states for 170 bp reference and mutant sequences; concatenate reference and mutant-minus-reference embeddings; fixed balanced L2 logistic head trained only on the MFASS training split. Rewire evaluation · Evaluation metadata: reproduced | ||
| 0.5500324040216661 auroc Unit: fraction · Direction: higher | Uncertainty: Point estimate; see paired-comparison artifacts. Coverage (scored/eligible): 8324/8324 | reproducedMFASS v2 pinned rewire artifacts · benchmarks/mfass/results/dnabert2-117m-frozen-pair-logreg.json :: auroc Source checking is not independent reproduction. |
| 0.04508654312652131 average_precision Unit: fraction · Direction: higher | Uncertainty: Point estimate; see paired-comparison artifacts. Coverage (scored/eligible): 8324/8324 | reproducedMFASS v2 pinned rewire artifacts · benchmarks/mfass/results/dnabert2-117m-frozen-pair-logreg.json :: average_precision Source checking is not independent reproduction. |
| 0.03 precision_at_100 Unit: fraction · Direction: higher | Uncertainty: Point estimate; see paired-comparison artifacts. Coverage (scored/eligible): 8324/8324 | reproducedMFASS v2 pinned rewire artifacts · benchmarks/mfass/results/dnabert2-117m-frozen-pair-logreg.json :: precision_at_100 Source checking is not independent reproduction. |
Inspected the pinned runner README and result configuration with git show. No models were run; existing result and reproduction statuses are unchanged.
Stable record: rewire-model-dnabert2-117m-frozen-pair-logregRelease 2026-09-16-d74d282221a9 · Record review: source checked
Stable ID: rewire-model-dnabert2-117m-frozen-pair-logreg