AUROC (cancer vs non-cancer, held-out ULP-WGS validation set) of Subnucleosome features with XGBoost (Fragmentia-AI paper re-implementation) on Geneseeq pan-cancer 0.1x ULP-WGS plasma cohort (2,004 cancers, 1,245 non-cancer)
0.676 AUROC (cancer vs non-cancer, held-out ULP-WGS validation set)
Methods
- Tested configuration
- Subnucleosome features with XGBoost (Fragmentia-AI paper re-implementation)
- Protocol
- Fragmentia-AI paper: ULP-WGS cancer vs non-cancer AUC on a held-out validation set, six feature sets x three classifiers
- Dataset
- Geneseeq pan-cancer 0.1x ULP-WGS plasma cohort (2,004 cancers, 1,245 non-cancer)
- Procedure
- Fragmentia-AI paper: ULP-WGS cancer vs non-cancer AUC on a held-out validation set, six feature sets x three classifiers
- Evaluation
- Subnucleosome / XGBoost on the ULP-WGS held-out validation set (Fragmentia-AI paper)
- Coverage
- Not reported scored / Not reported eligible
- Uncertainty
- Not reported by the source
- Evidence
- Independent external evaluation · source checkedToward generalizable prediction of cancer signal using a cell-free DNA language model; Fragmentia-AI paper, Document S1 (Figures S1-S9 and Tables S1-S6) · Table S5, row 'Subnucleosome' / 'XGBoost', column 'Performance (AUC)'
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
- Not reported
- Adaptation
- Trained on the ULP-WGS training cohort
- 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.
2 evidence rows matching the loaded filters
| Property and statement | Original source and location | Review and provenance |
|---|---|---|
| Reported result 0.6756 Individual claims | Toward generalizable prediction of cancer signal using a cell-free DNA language model Table S5, row 'Subnucleosome' / 'XGBoost', column 'Performance (AUC)' Shared locator for this statement’s cited sources; not a separate locator for each citation. Version: Cell Reports Medicine, published 2026-06-12; Europe PMC full-text JATS XML of PMC13400139 | source checked ["source-hash-verification","pdf-text-parse","independent-cell-check"] · 2026-10-12 independent paper Audit detailsRe-downloaded the pinned artifacts and matched their SHA-256 (the mmc1.pdf member of the Europe PMC bundle); compared all 19 Table S5 AUCs and 18 p values with a table transcribed for this review from the layout text (the extractor's script was not run). Field: Source artifact SHA-256: Hash scope: Hash scope not separately documented; inspect source record Extraction artifact SHA-256: |
| Reported result 0.6756 Individual claims | Fragmentia-AI paper, Document S1 (Figures S1-S9 and Tables S1-S6) Table S5, row 'Subnucleosome' / 'XGBoost', column 'Performance (AUC)' Shared locator for this statement’s cited sources; not a separate locator for each citation. Version: mmc1.pdf (Document S1) from the Europe PMC supplementary files of PMC13400139 | source checked ["source-hash-verification","pdf-text-parse","independent-cell-check"] · 2026-10-12 independent paper Audit detailsRe-downloaded the pinned artifacts and matched their SHA-256 (the mmc1.pdf member of the Europe PMC bundle); compared all 19 Table S5 AUCs and 18 p values with a table transcribed for this review from the layout text (the extractor's script was not run). Field: Source artifact SHA-256: Hash scope: artifact_sha256 is the mmc1.pdf member of the zip bundle served by Europe PMC; the bundle's own bytes vary between downloads Archive member: mmc1.pdf Extraction artifact SHA-256: |
Sources and history
Release 2026-11-21-c8bea25e5abb · Record review: source checked
2 source records and release history
- Toward generalizable prediction of cancer signal using a cell-free DNA language model · Original source · Cell Reports Medicine, published 2026-06-12; Europe PMC full-text JATS XML of PMC13400139
- Fragmentia-AI paper, Document S1 (Figures S1-S9 and Tables S1-S6) · Original source · mmc1.pdf (Document S1) from the Europe PMC supplementary files of PMC13400139
Technical metadata and extraction receipts
Stable ID: ctdnalit-20261011-fragmentia2026-result-subnucleosome-xgboost-auroc
- metric
- auroc
- metric qualifier
- cancer vs non-cancer, held-out ULP-WGS validation set
- metric direction
- higher
- unit
- unitless
- printed value
- 0.6756
- numeric value
- 0.6756
- source locator
- Table S5, row 'Subnucleosome' / 'XGBoost', column 'Performance (AUC)'
- review
- method: source-hash-verification; pdf-text-parse; independent-cell-check; reviewer: claude; reviewer note: Separate Claude review agent (reviewer-ctdna), independent of the extractor uc-ctdna; no human review claimed; date: 2026-10-12; artifact sha256: 357004eed301b424867442bb0c27b8ab94cc60e9a39dc4fc31e5728958122d41; retrieval url: https://www.ebi.ac.uk/europepmc/webservices/rest/PMC13400139/supplementaryFiles?includeInlineImage=false; method note: Extracted by deterministic parse of the PDF text layer (pdftotext -layout) with caption, header words, feature and classifier labels and row count asserted.; note: Re-downloaded the pinned artifacts and matched their SHA-256 (the mmc1.pdf member of the Europe PMC bundle); compared all 19 Table S5 AUCs and 18 p values with a table transcribed for this review from the layout text (the extractor's script was not run).
- missing metadata
- uncertainty: reason: unreported
- reported p value
- <0.0001, against Fragmentia-AI (Table S5 column 'P value (Compared with Fragmentia-AI)'; test not named)