rewirebio.iobenchmarks
Result

Hazard ratio (hazard ratio, higher-risk half against lower-risk half at the median predicted risk, progression-free survival, Cox stratified by treatment arm, Efron ties, full follow-up; equal-trial geometric mean over three trials) of BulkRNABert, published Hugging Face checkpoint (Ladon report comparator) on Three held-out trials with progression-free survival

1.6 Hazard ratio (hazard ratio, higher-risk half against lower-risk half at the median predicted risk, progression-free survival, Cox stratified by treatment arm, Efron ties, full follow-up; equal-trial geometric mean over three trials)

Methods

Tested configuration
BulkRNABert, published Hugging Face checkpoint (Ladon report comparator)
Protocol
Held-out trials, PFS hazard ratio of higher- against lower-risk half, geometric mean over three trials
Dataset subset
Three held-out trials with progression-free survival
Procedure
Held-out trials, PFS hazard ratio of higher- against lower-risk half, geometric mean over three trials
Evaluation
BulkRNABert, held-out trials PFS hazard ratio
Coverage
3/3 trials scored (All trials estimable.)
Uncertainty
Not reported by the source: Per-trial dots and a bar only; no interval is printed.
Evidence
Author-reported evaluation · source checkedLadon: An RNA Foundation Model for Predicting Patient Outcomes (Zenodo PDF) · Figure 3 (page 7), row 'BulkRNABert', label 'PFS 1.60'

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
Held-out trial; predictor fitted on TCGA and the other trials
Adaptation
BulkRNABert embeddings, ridge Cox, lambda = 0.3
Scoring implementation
Median split of predicted risk; unpenalised Cox with binary group indicator stratified by arm, Efron ties

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
1.60
Individual claims
Ladon: An RNA Foundation Model for Predicting Patient Outcomes (Zenodo PDF)

Original source ↗

Figure 3 (page 7), row 'BulkRNABert', label 'PFS 1.60'

Version: Zenodo record 24157116 (version DOI 10.5281/zenodo.24157116; concept DOI 10.5281/zenodo.24157115), file Ladon-An-RNA-Foundation-Model-for-Predicting-Patient-Outcomes.pdf, 47 pages, dated 12 November 2026
Retrieved: 2026-11-15T20:29:10Z

source checked

["independent-cell-check","pdf-text-parse","source-hash-verification"] · 2026-11-15T23:30:00Z

author reported

Audit details

Printed label re-read from the pinned PDF by coordinates; it matched.

Field: attributes.printed_value

Source artifact SHA-256: 64edf7c3c956c50a65aa01a0dcaa13f0cfef28ab4373ecfd50c431e67143893c

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

Extraction artifact SHA-256: 64edf7c3c956c50a65aa01a0dcaa13f0cfef28ab4373ecfd50c431e67143893c

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: ladon-20261115-result-fig3-bulkrnabert-pfs

areas
rna-transcriptomes
contexts
clinical_research
coverage
eligible: 3; note: All trials estimable.; scored: 3; unit: trials
metric
hazard-ratio
metric direction
higher
metric qualifier
hazard ratio, higher-risk half against lower-risk half at the median predicted risk, progression-free survival, Cox stratified by treatment arm, Efron ties, full follow-up; equal-trial geometric mean over three trials
missing metadata
uncertainty: note: Per-trial dots and a bar only; no interval is printed.; reason: unreported
numeric value
1.60
printed value
1.60
review
artifact sha256: 64edf7c3c956c50a65aa01a0dcaa13f0cfef28ab4373ecfd50c431e67143893c; method: independent-cell-check; pdf-text-parse; source-hash-verification; method note: Independent review pass. The Zenodo PDF was re-downloaded and its SHA-256 and MD5 checked against the source record and the Zenodo file checksum; the three HTML sources were refetched and all three hashes matched. The report prints no tables, so npm run batch -- check-values cannot place any locator and reports locator_not_parsed for all 101 results; value-check.tsv is kept as that record. All 101 were therefore re-read by review/check_figure_labels.py, written for this pass and using neither of the extractor's scripts: it reads the word-level geometry of a pdftotext -bbox-layout run and places every cell by its x and y coordinates, so each value is tied to the row whose label it sits beside and to the column or panel whose printed header it sits under, rather than to a line of a -layout text layer or a character offset. Figure 2's three panels and Figure 17's two columns were resolved by clustering the value positions and ordering the clusters by the printed headers; every printed 95% interval was matched to the result's uncertainty bounds. 113 checks, no disagreement. Three arithmetic constraints were re-derived: Figure 17's five per-lineage Pearson r values average to the bottom chart's overall r for both Ladon (0.872 against a printed 0.87) and CIBERSORT (0.594 against 0.600), which independently fixes which of the two numbers in each lineage label is Ladon's; Figure 2's pan-cancer bar differs from the mean of its two transfer bars, as the caption requires; and Figure 14's in-figure top-5% label agrees with the page 27 prose. A manual sample was read by eye from the -layout text layer: all 24 Figure 2 cells, all 14 Figure 3 OS and PFS values, all 10 Figure 18 AUCs with their intervals, and the Figure 5, 11, 12a, 13 and 17 readouts, with column and panel assignment checked in each. Figure 2's 24 cells, Figure 3's 14 and Figure 18's 10 AUCs and intervals were confirmed a third time against the facet and row keys of the authors' own figure pages.; note: Printed label re-read from the pinned PDF by coordinates; it matched.; retrieval url: https://zenodo.org/api/records/24157116/files/Ladon-An-RNA-Foundation-Model-for-Predicting-Patient-Outcomes.pdf/content; reviewed at: 2026-11-15T23:30:00Z; reviewer: claude; reviewer note: Claude review agent, independent of the extractor; no human review claimed. Transcription was checked against the pinned sources, which is not reproduction: nothing was rerun and no value was recomputed.
source locator
Figure 3 (page 7), row 'BulkRNABert', label 'PFS 1.60'
unit
unitless
Related records

Suggest a correction