rewirebio.iobenchmarks
Result

Accuracy (gene region class ncRNA, leave-one-chromosome-out logistic regression on averaged 100 bp GPN embeddings) of GPN trained on eight Brassicales genomes, repeat loss weight 0.1 (PNAS 2023) on Non-overlapping 100 bp windows of the Arabidopsis thaliana genome, labelled by gene region

51% Accuracy (gene region class ncRNA, leave-one-chromosome-out logistic regression on averaged 100 bp GPN embeddings)

Methods

Tested configuration
GPN trained on eight Brassicales genomes, repeat loss weight 0.1 (PNAS 2023)
Protocol
Gene region classification from averaged GPN embeddings, leave-one-chromosome-out
Dataset subset
Non-overlapping 100 bp windows of the Arabidopsis thaliana genome, labelled by gene region
Procedure
Gene region classification from averaged GPN embeddings, leave-one-chromosome-out
Evaluation
Gene region classification from averaged GPN embeddings (PNAS 2023)
Coverage
Not reported scored / Not reported eligible
Uncertainty
Not reported by the source: The source prints no interval or spread for this value.
Evidence
Author-reported evaluation · source checkedDNA language models are powerful predictors of genome-wide variant effects (PNAS 2023) · Results, 'Unsupervised Clustering of Genomic Regions', sentence beginning "To quantify GPN's ability to distinguish genomic regions", term "ncRNA (51%)"

A source-checked result verifies the numerical transcription, not every model or protocol detail. Evaluation metadata: source checked. Source checked does not mean independently reproduced.

Reproduction

Split
Leave one chromosome out: each chromosome predicted by a model trained on the remaining chromosomes
Adaptation
Supervised logistic regression probe, class weight inversely proportional to frequency, L2 strength chosen by cross-validation (scikit-learn)
Scoring implementation
Not reported

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.

1 evidence row matching the loaded filters

Claims, original sources and review scope · Release 2026-11-21-c8bea25e5abb
Property and statementOriginal source and locationReview and provenance
Reported result
51%
Individual claims
DNA language models are powerful predictors of genome-wide variant effects (PNAS 2023)

Original source ↗

Results, 'Unsupervised Clustering of Genomic Regions', sentence beginning "To quantify GPN's ability to distinguish genomic regions", term "ncRNA (51%)"

Version: PNAS 120(44) e2311219120, published 26 October 2023; Europe PMC full text JATS XML of PMC10622914
Retrieved: 2026-11-16T00:29:00Z

source checked

["independent-cell-check","automated-source-review","source-hash-verification"] · 2026-11-16T01:30:00Z

author reported

Audit details

'the lowest on ncRNA (51%), the least frequent class' upheld on the same reading as the CDS result.

Field: attributes.printed_value

Source artifact SHA-256: 8ff2184c07f5e6807af0b9d5103d669e1920e735ab5a4b1ea856c070077ae1e0

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

Format: JATS XML

Extraction artifact SHA-256: 8ff2184c07f5e6807af0b9d5103d669e1920e735ab5a4b1ea856c070077ae1e0

Extraction artifact

Sources and history

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

1 source record and release historyDownload this release (gzip)
Technical metadata and extraction receipts

Stable ID: gpn-paper-20261116-result-gene-region-accuracy-ncrna

areas
dna-genomes
contexts
research
metric
accuracy
metric direction
higher
metric qualifier
gene region class ncRNA, leave-one-chromosome-out logistic regression on averaged 100 bp GPN embeddings
missing metadata
uncertainty: note: The source prints no interval or spread for this value.; reason: unreported
numeric value
51
printed value
51%
review
artifact sha256: 8ff2184c07f5e6807af0b9d5103d669e1920e735ab5a4b1ea856c070077ae1e0; method: independent-cell-check; automated-source-review; source-hash-verification; method note: Independent review pass by a different agent from the extractor, with its own script (review/recheck_values.py), which does not import extract/parse_values.py and does not read extract/values.json. The published article XML and the bioRxiv details API response were re-downloaded and matched their pinned hashes; the SI Appendix PDF was re-downloaded as the member of a freshly built Europe PMC zip (a third distinct zip hash, c7a1b20e8796c4bd384ba6f0308134e32a66e40dc881b0c1e3d0201da26e8498) and the PDF member matched. Every archived gz in artifacts/ decompresses to the hash its source record pins, and the text layer in extract/ regenerates byte for byte from the archived PDF with pdftotext -layout, poppler 26.08.0. npm run batch -- check-values reproduced the extractor's figures: 6 in_cell, 17 locator_not_parsed, no overclaimed_in_row. The article was parsed with xml.etree, the section tree read from its own <sec><title> elements, and each value checked inside the section its locator names: the value together with the label the locator names must occur exactly once in that section and exactly once in the whole body, and a negative control with the value moved to a neighbouring label must be absent. Whitespace is dropped before comparison, because the JATS marks up italics, maths and superscripts inside these sentences, so the spacing is an artefact of the markup and not of the page. A mutation test (review/mutation_test.py) confirmed the check fails on a changed value, including the one value the bioRxiv v3 prints differently.; note: 'the lowest on ncRNA (51%), the least frequent class' upheld on the same reading as the CDS result.; retrieval url: https://www.ebi.ac.uk/europepmc/webservices/rest/PMC10622914/fullTextXML; reviewed at: 2026-11-16T01:30:00Z; reviewer: claude; reviewer note: Claude review agent, independent of the extractor; no human review claimed. Transcription was checked against the pinned artifacts; nothing was reproduced and no value was recomputed.
scope note
The text calls this the accuracy on this class. Whether the printed quantity is per-class recall is not stated.
source locator
Results, 'Unsupervised Clustering of Genomic Regions', sentence beginning "To quantify GPN's ability to distinguish genomic regions", term "ncRNA (51%)"
unit
percent
Related records

Suggest a correction