rewirebio.iobenchmarks
Evaluation

Nucleotide Transformer on ClinVar Splice Donor variants, likelihood-ratio scoring

Author-reported comparison transcribed from the pinned preprint; not reproduced.

Evaluation results

1 evaluation · 1 result. 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: Nucleotide Transformer v2 500M multi-species, likelihood-ratio scoringProtocol: ClinVar pathogenic versus benign classification from likelihood ratios, AUROC by variant class
Dataset subset: ClinVar Splice Donor variants, at least one review star
0.67 AUROC (ClinVar pathogenic or likely pathogenic versus benign or likely benign, zero-shot allele log likelihood ratio, class Splice Donor)
unitless · higher

Uncertainty: Not reported by the source: Table 1 prints no interval or spread.

Coverage: Not reported scored / Not reported eligible

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

Nucleotide Transformer on ClinVar Splice Donor variants, likelihood-ratio scoring

phylogpn-20261116-protocol-clinvar-llr-auroc

Aggregation: Not reported

A Phylogenetic Approach to Genomic Language Modeling · Table 1, row 'Splice Donor', column 'Nucleotide Transformer'

Source checking is not independent reproduction. Release 2026-11-21-c8bea25e5abb.

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-c8bea25e5abb · 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.

Evaluation procedure

ClinVar pathogenic versus benign classification from likelihood ratios, AUROC by variant class

Configuration
Nucleotide Transformer v2 500M multi-species, likelihood-ratio scoring
Protocol
ClinVar pathogenic versus benign classification from likelihood ratios, AUROC by variant class
Dataset subset
ClinVar Splice Donor variants, at least one review star
origin
Author-reported evaluation
configuration
Not reported
protocol id
phylogpn-20261116-protocol-clinvar-llr-auroc
dataset version
Not reported
split
None: zero-shot likelihood-ratio scoring
population
ClinVar variants with at least one review star, class Splice Donor
inputs
At least 9,283 bp of sequence context per scored position
adaptation
None: 6-mer token likelihood ratio averaged with the reverse complement
metric implementation
AUROC with pathogenic and likely pathogenic variants as positives
aggregation
One AUROC over the variants of this class
budget
Not reported

Metadata review: source checked. Unreported conditions prevent automatic comparisons.

Reproduction

Split
None: zero-shot likelihood-ratio scoring
Adaptation
None: 6-mer token likelihood ratio averaged with the reverse complement
Scoring implementation
AUROC with pathogenic and likely pathogenic variants as positives

No execution recipe has been verified for this exact configuration and evaluation. A benchmark's general instructions may use different inputs, splits or model settings.

Reproducing this published result requires matching its model configuration, data, split and scorer. Source checking or a successful smoke test does not establish score reproduction.

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.

34 evidence rows matching the loaded filters

Claims, original sources and review scope · Release 2026-11-21-c8bea25e5abb
Property and statementOriginal source and locationReview and provenance
Comparison: adaptation
None: 6-mer token likelihood ratio averaged with the reverse complement
Context-only references
A Phylogenetic Approach to Genomic Language Modeling

Original source ↗

Table 1, row 'Splice Donor', column 'Nucleotide Transformer'

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

Version: preprint version in PMC
Retrieved: 2026-09-16T10:38:57.558218+00:00

not individually reviewed

No individual claim review recorded

author reported

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

Audit details

Field: attributes.comparison.adaptation

Source artifact SHA-256: 807f3a26cbfa9b5ce238d92164bd523302c67d1c5794b08273c51cca1acd4224

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

Inspected artifact

Comparison: adaptation
None: 6-mer token likelihood ratio averaged with the reverse complement
Context-only references
Supplement 1 of the PhyloGPN preprint (arXiv:2503.03773v1)

Original source ↗

Table 1, row 'Splice Donor', column 'Nucleotide Transformer'

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

Version: Supplement 1 of arXiv:2503.03773v1 as deposited in PMC11908359
Retrieved: 2026-11-16T01:07:00Z

not individually reviewed

No individual claim review recorded

author reported

Audit details

Field: attributes.comparison.adaptation

Source artifact SHA-256: b4885d288e03ae976acafb2b6b1c3837572dce224c8803715f4d329966ba65f4

Hash scope: SHA-256 of the supplement PDF taken from the Europe PMC supplementaryFiles zip for PMC11908359. The zip is not byte-stable: two retrievals on 2026-11-16 gave different bytes because the archive carries per-download timestamps, while the member PDF was byte-identical in both.

Format: application/pdf

Archive member: NIHPP2503.03773v1-supplement-1.pdf

Inspected artifact

Comparison: aggregation
One AUROC over the variants of this class
Context-only references
A Phylogenetic Approach to Genomic Language Modeling

Original source ↗

Table 1, row 'Splice Donor', column 'Nucleotide Transformer'

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

Version: preprint version in PMC
Retrieved: 2026-09-16T10:38:57.558218+00:00

not individually reviewed

No individual claim review recorded

author reported

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

Audit details

Field: attributes.comparison.aggregation

Source artifact SHA-256: 807f3a26cbfa9b5ce238d92164bd523302c67d1c5794b08273c51cca1acd4224

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

Inspected artifact

Comparison: aggregation
One AUROC over the variants of this class
Context-only references
Supplement 1 of the PhyloGPN preprint (arXiv:2503.03773v1)

Original source ↗

Table 1, row 'Splice Donor', column 'Nucleotide Transformer'

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

Version: Supplement 1 of arXiv:2503.03773v1 as deposited in PMC11908359
Retrieved: 2026-11-16T01:07:00Z

not individually reviewed

No individual claim review recorded

author reported

Audit details

Field: attributes.comparison.aggregation

Source artifact SHA-256: b4885d288e03ae976acafb2b6b1c3837572dce224c8803715f4d329966ba65f4

Hash scope: SHA-256 of the supplement PDF taken from the Europe PMC supplementaryFiles zip for PMC11908359. The zip is not byte-stable: two retrievals on 2026-11-16 gave different bytes because the archive carries per-download timestamps, while the member PDF was byte-identical in both.

Format: application/pdf

Archive member: NIHPP2503.03773v1-supplement-1.pdf

Inspected artifact

Comparison: budget
Not reported
Context-only references
A Phylogenetic Approach to Genomic Language Modeling

Original source ↗

Table 1, row 'Splice Donor', column 'Nucleotide Transformer'

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

Version: preprint version in PMC
Retrieved: 2026-09-16T10:38:57.558218+00:00

missing or unspecified

No individual claim review recorded

author reported

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

Audit details

Field: attributes.comparison.budget

Source artifact SHA-256: 807f3a26cbfa9b5ce238d92164bd523302c67d1c5794b08273c51cca1acd4224

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

Inspected artifact

Comparison: budget
Not reported
Context-only references
Supplement 1 of the PhyloGPN preprint (arXiv:2503.03773v1)

Original source ↗

Table 1, row 'Splice Donor', column 'Nucleotide Transformer'

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

Version: Supplement 1 of arXiv:2503.03773v1 as deposited in PMC11908359
Retrieved: 2026-11-16T01:07:00Z

missing or unspecified

No individual claim review recorded

author reported

Audit details

Field: attributes.comparison.budget

Source artifact SHA-256: b4885d288e03ae976acafb2b6b1c3837572dce224c8803715f4d329966ba65f4

Hash scope: SHA-256 of the supplement PDF taken from the Europe PMC supplementaryFiles zip for PMC11908359. The zip is not byte-stable: two retrievals on 2026-11-16 gave different bytes because the archive carries per-download timestamps, while the member PDF was byte-identical in both.

Format: application/pdf

Archive member: NIHPP2503.03773v1-supplement-1.pdf

Inspected artifact

Comparison: dataset version
Not reported
Context-only references
A Phylogenetic Approach to Genomic Language Modeling

Original source ↗

Table 1, row 'Splice Donor', column 'Nucleotide Transformer'

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

Version: preprint version in PMC
Retrieved: 2026-09-16T10:38:57.558218+00:00

missing or unspecified

No individual claim review recorded

author reported

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

Audit details

Field: attributes.comparison.dataset_version

Source artifact SHA-256: 807f3a26cbfa9b5ce238d92164bd523302c67d1c5794b08273c51cca1acd4224

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

Inspected artifact

Comparison: dataset version
Not reported
Context-only references
Supplement 1 of the PhyloGPN preprint (arXiv:2503.03773v1)

Original source ↗

Table 1, row 'Splice Donor', column 'Nucleotide Transformer'

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

Version: Supplement 1 of arXiv:2503.03773v1 as deposited in PMC11908359
Retrieved: 2026-11-16T01:07:00Z

missing or unspecified

No individual claim review recorded

author reported

Audit details

Field: attributes.comparison.dataset_version

Source artifact SHA-256: b4885d288e03ae976acafb2b6b1c3837572dce224c8803715f4d329966ba65f4

Hash scope: SHA-256 of the supplement PDF taken from the Europe PMC supplementaryFiles zip for PMC11908359. The zip is not byte-stable: two retrievals on 2026-11-16 gave different bytes because the archive carries per-download timestamps, while the member PDF was byte-identical in both.

Format: application/pdf

Archive member: NIHPP2503.03773v1-supplement-1.pdf

Inspected artifact

Comparison: inputs
At least 9,283 bp of sequence context per scored position
Context-only references
A Phylogenetic Approach to Genomic Language Modeling

Original source ↗

Table 1, row 'Splice Donor', column 'Nucleotide Transformer'

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

Version: preprint version in PMC
Retrieved: 2026-09-16T10:38:57.558218+00:00

not individually reviewed

No individual claim review recorded

author reported

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

Audit details

Field: attributes.comparison.inputs

Source artifact SHA-256: 807f3a26cbfa9b5ce238d92164bd523302c67d1c5794b08273c51cca1acd4224

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

Inspected artifact

Comparison: inputs
At least 9,283 bp of sequence context per scored position
Context-only references
Supplement 1 of the PhyloGPN preprint (arXiv:2503.03773v1)

Original source ↗

Table 1, row 'Splice Donor', column 'Nucleotide Transformer'

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

Version: Supplement 1 of arXiv:2503.03773v1 as deposited in PMC11908359
Retrieved: 2026-11-16T01:07:00Z

not individually reviewed

No individual claim review recorded

author reported

Audit details

Field: attributes.comparison.inputs

Source artifact SHA-256: b4885d288e03ae976acafb2b6b1c3837572dce224c8803715f4d329966ba65f4

Hash scope: SHA-256 of the supplement PDF taken from the Europe PMC supplementaryFiles zip for PMC11908359. The zip is not byte-stable: two retrievals on 2026-11-16 gave different bytes because the archive carries per-download timestamps, while the member PDF was byte-identical in both.

Format: application/pdf

Archive member: NIHPP2503.03773v1-supplement-1.pdf

Inspected artifact

How to read the evidence

Sources and history

Release 2026-11-21-c8bea25e5abb · Record review: source checked

2 source records and release historyDownload this release (gzip)
Technical metadata and extraction receipts

Stable ID: phylogpn-20261116-eval-t1-splice-donor-nucleotide-transformer

areas
dna-genomes
contexts
clinical_research
tasks
ClinVar variant classification
origin
author_reported
protocol
phylogpn-20261116-protocol-clinvar-llr-auroc
comparison
protocol id: phylogpn-20261116-protocol-clinvar-llr-auroc; dataset version: Not reported; split: None: zero-shot likelihood-ratio scoring; population: ClinVar variants with at least one review star, class Splice Donor; inputs: At least 9,283 bp of sequence context per scored position; adaptation: None: 6-mer token likelihood ratio averaged with the reverse complement; metric implementation: AUROC with pathogenic and likely pathogenic variants as positives; aggregation: One AUROC over the variants of this class; budget: Not reported
source locator
Table 1, row 'Splice Donor', column 'Nucleotide Transformer'
missing metadata
comparison.budget: reason: inapplicable; note: Zero-shot or precomputed scoring; no compute budget applies.; comparison.dataset version: reason: unreported; note: No ClinVar release or download date is printed.
review
method: independent-cell-check; deterministic-table-parse; source-hash-verification; method note: Independent review pass by a different agent from the extractor, on the extraction frozen at batch.jsonl SHA-256 fbf8f6410ec0063a5e7beea4b0c8842052cafe5d26be31d70cbe9e341dce4903. The pinned Europe PMC full text XML was re-retrieved and its SHA-256 matched the stored source record byte for byte. All 124 results were re-read by review/reread.py, written for this pass, which parses the two table-wrap elements with xml.etree, keys every cell by its printed row label and column header, and does not use extract/parse_tables.py or extract/tables.json. For each result it checks the printed value, the numeric value, the metric concept against the column header's own parenthetical (MCC, AUPRC, AUROC) or the Table 1 caption, the direction, the unit, that uncertainty and denominator are recorded as unreported, that each locator resolves to exactly one cell, that no cell is claimed twice and that every printed cell has exactly one result; and it checks each evaluation's row-to-configuration assignment against a mapping written independently from the paper. review/controls.py shows the check rejects a swapped value, a value shifted one column, a row label shifted one row, and a zero recorded for a printed N/A cell. The six printed N/A cells were read in the raw markup; no result, no evaluation and no zero stands in for them. npm run batch -- check-values reported 124 in_cell and reproduced value-check.tsv byte for byte. A manual sample of ten rows, 70 cells, including the three asterisked cells and one row from each family-best row, was read by eye in the raw markup (review/manual-sample.txt). review/family-best.py recomputed all 35 compiled Table 2 cells against the BEND results already in the store.; reviewer: claude; reviewer note: Claude review agent (independent of the extractor); no human review claimed. Transcription was checked against the pinned sources; nothing was reproduced.; date: 2026-11-16; artifact sha256: 807f3a26cbfa9b5ce238d92164bd523302c67d1c5794b08273c51cca1acd4224; retrieval url: https://www.ebi.ac.uk/europepmc/webservices/rest/PMC11908359/fullTextXML; note: Cell re-read at Table 1, row 'Splice Donor', column 'Nucleotide Transformer' in the pinned JATS XML by review/reread.py; the printed value, the numeric value, the metric taken from the printed column header, the direction, the unit and the row-to-configuration assignment all matched.
Related records

Suggest a correction