rewirebio.iobenchmarks
Source

Benchmarking DNA Foundation Models: Biological Blind Spots in Evo2 Variant-Effect Prediction

bioRxiv preprint version 1, posted 2026-03-11 · Preprint

DOI: 10.64898/2026.03.10.710786

What this source contributed

Records in this release that cite this source. A record can cite several sources.

Results

Values as the source printed them. Compare results only under compatible protocols, datasets and metrics.

Model or configurationMetricValueDatasetEvaluation
Evo2 (Youden) (Mathur and Sachidanandam 2026)Accuracy (Youden threshold from calibration set)0.832 fractionMitochondrial variant test split, MITOMAP pathogenic and ClinGen/ClinVar benign (Mathur and Sachidanandam 2026)Evo2 (Youden) on mtDNA pathogenicity test split
Evo2 (Youden) (Mathur and Sachidanandam 2026)AUPRC0.674 unitlessMitochondrial variant test split, MITOMAP pathogenic and ClinGen/ClinVar benign (Mathur and Sachidanandam 2026)Evo2 (Youden) on mtDNA pathogenicity test split
Evo2 (Youden) (Mathur and Sachidanandam 2026)AUROC0.896 unitlessMitochondrial variant test split, MITOMAP pathogenic and ClinGen/ClinVar benign (Mathur and Sachidanandam 2026)Evo2 (Youden) on mtDNA pathogenicity test split
Evo2 (Youden) (Mathur and Sachidanandam 2026)Balanced accuracy (Youden threshold from calibration set)0.846 fractionMitochondrial variant test split, MITOMAP pathogenic and ClinGen/ClinVar benign (Mathur and Sachidanandam 2026)Evo2 (Youden) on mtDNA pathogenicity test split
Evo2 (Youden) (Mathur and Sachidanandam 2026)MCC (Youden threshold from calibration set)0.631 unitlessMitochondrial variant test split, MITOMAP pathogenic and ClinGen/ClinVar benign (Mathur and Sachidanandam 2026)Evo2 (Youden) on mtDNA pathogenicity test split
Evo2 (Youden) (Mathur and Sachidanandam 2026)Precision (Youden threshold from calibration set)0.500 fractionMitochondrial variant test split, MITOMAP pathogenic and ClinGen/ClinVar benign (Mathur and Sachidanandam 2026)Evo2 (Youden) on mtDNA pathogenicity test split
Evo2 (Youden) (Mathur and Sachidanandam 2026)Recall (Youden threshold from calibration set)0.867 fractionMitochondrial variant test split, MITOMAP pathogenic and ClinGen/ClinVar benign (Mathur and Sachidanandam 2026)Evo2 (Youden) on mtDNA pathogenicity test split
Evo2 (Youden) (Mathur and Sachidanandam 2026)Specificity (Youden threshold from calibration set)0.825 fractionMitochondrial variant test split, MITOMAP pathogenic and ClinGen/ClinVar benign (Mathur and Sachidanandam 2026)Evo2 (Youden) on mtDNA pathogenicity test split

Models, benchmarks and other records

Configurations (1)

Protocols (1)

Datasets (1)

Evaluations (1)

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
Artifact SHA-256
8fe2977a11335823be19474b350b65d6d436c7674e92e96f81c9086df0dacf7d
Source metadata
Benchmarking DNA Foundation Models: Biological Blind Spots in Evo2 Variant-Effect Prediction

Original source ↗

No field-specific location recorded

Version: bioRxiv preprint version 1, posted 2026-03-11
Retrieved: 2026-10-11T00:44:59Z

catalogued

No individual claim review recorded

Source has a recorded evidence concern. Consult its source page before using the claim.

Audit details

Field: attributes.artifact_sha256

Source artifact SHA-256: 8fe2977a11335823be19474b350b65d6d436c7674e92e96f81c9086df0dacf7d

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

Inspected artifact

Artifact url
https://www.biorxiv.org/content/10.64898/2026.03.10.710786v1.full.pdf
Source metadata
Benchmarking DNA Foundation Models: Biological Blind Spots in Evo2 Variant-Effect Prediction

Original source ↗

No field-specific location recorded

Version: bioRxiv preprint version 1, posted 2026-03-11
Retrieved: 2026-10-11T00:44:59Z

catalogued

No individual claim review recorded

Source has a recorded evidence concern. Consult its source page before using the claim.

Audit details

Field: attributes.artifact_url

Source artifact SHA-256: 8fe2977a11335823be19474b350b65d6d436c7674e92e96f81c9086df0dacf7d

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

Inspected artifact

Doi
10.64898/2026.03.10.710786
Source metadata
Benchmarking DNA Foundation Models: Biological Blind Spots in Evo2 Variant-Effect Prediction

Original source ↗

No field-specific location recorded

Version: bioRxiv preprint version 1, posted 2026-03-11
Retrieved: 2026-10-11T00:44:59Z

catalogued

No individual claim review recorded

Source has a recorded evidence concern. Consult its source page before using the claim.

Audit details

Field: attributes.doi

Source artifact SHA-256: 8fe2977a11335823be19474b350b65d6d436c7674e92e96f81c9086df0dacf7d

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

Inspected artifact

Evidence concerns
1 values
  • source id
    lit-genomics-20261011-evo2mito-source-mathur2026
    message
    Comparator rows reproduce APOGEE 2 (Bianco et al. 2023) cross-validation results on a missense-only set (140 pathogenic, 1,734 neutral). They were not measured on this paper's test split, so Table 2 does not compare like with like. Separately, Table 2 sensitivity/specificity 0.867/0.825 (balanced accuracy 0.846) versus Figure 2B and text 90.2%/85.0% (87.6%); the source does not explain the difference.
    source locator
    Section 3 Methods (comparison paragraph, reference 27); Table 2; Figure 2B legend; Section 4.3.1
    artifact sha256
    8fe2977a11335823be19474b350b65d6d436c7674e92e96f81c9086df0dacf7d
    review date
    2026-10-11
Source metadata
Benchmarking DNA Foundation Models: Biological Blind Spots in Evo2 Variant-Effect Prediction

Original source ↗

No field-specific location recorded

Version: bioRxiv preprint version 1, posted 2026-03-11
Retrieved: 2026-10-11T00:44:59Z

catalogued

No individual claim review recorded

Source has a recorded evidence concern. Consult its source page before using the claim.

Audit details

Field: attributes.evidence_concerns

Source artifact SHA-256: 8fe2977a11335823be19474b350b65d6d436c7674e92e96f81c9086df0dacf7d

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

Inspected artifact

Licence
CC-BY-4.0
Source metadata
Benchmarking DNA Foundation Models: Biological Blind Spots in Evo2 Variant-Effect Prediction

Original source ↗

No field-specific location recorded

Version: bioRxiv preprint version 1, posted 2026-03-11
Retrieved: 2026-10-11T00:44:59Z

catalogued

No individual claim review recorded

Source has a recorded evidence concern. Consult its source page before using the claim.

Audit details

Field: attributes.licence

Source artifact SHA-256: 8fe2977a11335823be19474b350b65d6d436c7674e92e96f81c9086df0dacf7d

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

Inspected artifact

Limitations
3 values
  • Preprint, not peer reviewed.
  • The Evo2 checkpoint (size) used is not stated.
  • Comparator rows of Table 2 reproduce APOGEE 2 (Bianco et al. 2023, Nat Commun 14:5058) cross-validation results on a missense-only set (140 pathogenic, 1,734 neutral). They were not measured on this paper's test split and are not extracted; only the Evo2 row is recorded.
Source metadata
Benchmarking DNA Foundation Models: Biological Blind Spots in Evo2 Variant-Effect Prediction

Original source ↗

No field-specific location recorded

Version: bioRxiv preprint version 1, posted 2026-03-11
Retrieved: 2026-10-11T00:44:59Z

catalogued

No individual claim review recorded

Source has a recorded evidence concern. Consult its source page before using the claim.

Audit details

Field: attributes.limitations

Source artifact SHA-256: 8fe2977a11335823be19474b350b65d6d436c7674e92e96f81c9086df0dacf7d

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

Inspected artifact

Media type
application/pdf
Source metadata
Benchmarking DNA Foundation Models: Biological Blind Spots in Evo2 Variant-Effect Prediction

Original source ↗

No field-specific location recorded

Version: bioRxiv preprint version 1, posted 2026-03-11
Retrieved: 2026-10-11T00:44:59Z

catalogued

No individual claim review recorded

Source has a recorded evidence concern. Consult its source page before using the claim.

Audit details

Field: attributes.media_type

Source artifact SHA-256: 8fe2977a11335823be19474b350b65d6d436c7674e92e96f81c9086df0dacf7d

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

Inspected artifact

Publication status
preprint
Source metadata
Benchmarking DNA Foundation Models: Biological Blind Spots in Evo2 Variant-Effect Prediction

Original source ↗

No field-specific location recorded

Version: bioRxiv preprint version 1, posted 2026-03-11
Retrieved: 2026-10-11T00:44:59Z

catalogued

No individual claim review recorded

Source has a recorded evidence concern. Consult its source page before using the claim.

Audit details

Field: attributes.publication_status

Source artifact SHA-256: 8fe2977a11335823be19474b350b65d6d436c7674e92e96f81c9086df0dacf7d

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

Inspected artifact

Retrieved at
2026-10-11T00:44:59Z
Source metadata
Benchmarking DNA Foundation Models: Biological Blind Spots in Evo2 Variant-Effect Prediction

Original source ↗

No field-specific location recorded

Version: bioRxiv preprint version 1, posted 2026-03-11
Retrieved: 2026-10-11T00:44:59Z

catalogued

No individual claim review recorded

Source has a recorded evidence concern. Consult its source page before using the claim.

Audit details

Field: attributes.retrieved_at

Source artifact SHA-256: 8fe2977a11335823be19474b350b65d6d436c7674e92e96f81c9086df0dacf7d

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

Inspected artifact

Review: date
2026-10-11
Source metadata
Benchmarking DNA Foundation Models: Biological Blind Spots in Evo2 Variant-Effect Prediction

Original source ↗

No field-specific location recorded

Version: bioRxiv preprint version 1, posted 2026-03-11
Retrieved: 2026-10-11T00:44:59Z

catalogued

No individual claim review recorded

Source has a recorded evidence concern. Consult its source page before using the claim.

Audit details

Field: attributes.review.date

Source artifact SHA-256: 8fe2977a11335823be19474b350b65d6d436c7674e92e96f81c9086df0dacf7d

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

Inspected artifact

How to read the evidence

Sources and history

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

Primary source retrieved and hashed for the 2026-10-11 genomics literature pass.

Release history

No supporting source is linked yet.

Read original source

Download this release (gzip)
Technical metadata and extraction receipts

Stable ID: lit-genomics-20261011-evo2mito-source-mathur2026

areas
dna-genomes
contexts
research
url
https://doi.org/10.64898/2026.03.10.710786
artifact url
https://www.biorxiv.org/content/10.64898/2026.03.10.710786v1.full.pdf
version
bioRxiv preprint version 1, posted 2026-03-11
retrieved at
2026-10-11T00:44:59Z
artifact sha256
8fe2977a11335823be19474b350b65d6d436c7674e92e96f81c9086df0dacf7d
doi
10.64898/2026.03.10.710786
publication status
preprint
licence
CC-BY-4.0
media type
application/pdf
limitations
Preprint, not peer reviewed.; The Evo2 checkpoint (size) used is not stated.; Comparator rows of Table 2 reproduce APOGEE 2 (Bianco et al. 2023, Nat Commun 14:5058) cross-validation results on a missense-only set (140 pathogenic, 1,734 neutral). They were not measured on this paper's test split and are not extracted; only the Evo2 row is recorded.
evidence concerns
source id: lit-genomics-20261011-evo2mito-source-mathur2026; message: Comparator rows reproduce APOGEE 2 (Bianco et al. 2023) cross-validation results on a missense-only set (140 pathogenic, 1,734 neutral). They were not measured on this paper's test split, so Table 2 does not compare like with like. Separately, Table 2 sensitivity/specificity 0.867/0.825 (balanced accuracy 0.846) versus Figure 2B and text 90.2%/85.0% (87.6%); the source does not explain the difference.; source locator: Section 3 Methods (comparison paragraph, reference 27); Table 2; Figure 2B legend; Section 4.3.1; artifact sha256: 8fe2977a11335823be19474b350b65d6d436c7674e92e96f81c9086df0dacf7d; review date: 2026-10-11
review
method: source-hash-verification; method note: Artifact re-retrieved or unpacked and SHA-256 matched.; reviewer: claude; reviewer note: Claude review agent reviewer-gen-1, independent of the extractor lit-genomics; no human review claimed; date: 2026-10-11
Related records

    Suggest a correction