Relevance of ctdnalit-20261011-cpgpt2026-protocol-dmr-informed-split-runs to "Which measured methylation workflow detects tumour-derived plasma DNA and, where separately supported, identifies tissue of origin?"
- Use case
- Select a plasma ctDNA methylation detection workflow
- Assessed by
- Xu and Feng 2026: GI cancer vs non-cancer on GSE149438, three train-validation-test split-runs, DMR-informed CpG features
- Relevance
- Proxy evidence
- Endpoint
- GI cancer vs non-cancer test AUROC and AUPRC (and validation AUROC), mean over three split-runs, for a fine-tuned methylation foundation model (CpGPT) and four classical classifiers on the same targeted bisulfite cfDNA features and partitions
Compares a methylation foundation model with conventional classifiers on the same plasma cfDNA methylation samples and splits. Proxy: AUROC without a declared specificity, very small test partitions, feature selection that saw the test samples, and an unreviewed preprint.
5 reviewed evaluations
- CpGPT on GSE149438 DMR-informed split-runs (Xu and Feng 2026)
- Elastic Net on GSE149438 DMR-informed split-runs (Xu and Feng 2026)
- GBM on GSE149438 DMR-informed split-runs (Xu and Feng 2026)
- Random Forest on GSE149438 DMR-informed split-runs (Xu and Feng 2026)
- SVM radial on GSE149438 DMR-informed split-runs (Xu and Feng 2026)
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.
18 evidence rows matching the loaded filters
| Property and statement | Original source and location | Review and provenance |
|---|---|---|
Citation locations1 values
| An Evaluation of DMR-Informed Fine-Tuning of Tissue Array-Pretrained CpGPT for Gastrointestinal Cancer Classification Using cfDNA Targeted Methylation ctdnalit-20261011-cpgpt2026-source-preprint: Table 3; Methods; Results Version: medRxiv version 1, posted 2026-07-29 | not individually reviewed No individual claim review recorded Audit detailsField: Source artifact SHA-256: Hash scope: artifact_sha256 describes one retrieval (2026-10-11T21:49:59Z). medRxiv regenerates this PDF: a reviewer download on 2026-10-12 had 2,761,577 bytes, SHA-256 3d37b32e8ff782201ed8ee6677a01808d4f8c86eea70350a530dde7306cab7c9 and PDF CreationDate 2026-10-11T20:08:12-07:00. The reviewer's copy's text layer matched every extracted value. The licence does not allow archiving. |
| Comparison group cpgpt2026-gse149438 Context-only references | An Evaluation of DMR-Informed Fine-Tuning of Tissue Array-Pretrained CpGPT for Gastrointestinal Cancer Classification Using cfDNA Targeted Methylation ctdnalit-20261011-cpgpt2026-source-preprint: Table 3; Methods; Results Version: medRxiv version 1, posted 2026-07-29 | not individually reviewed No individual claim review recorded Audit detailsField: Source artifact SHA-256: Hash scope: artifact_sha256 describes one retrieval (2026-10-11T21:49:59Z). medRxiv regenerates this PDF: a reviewer download on 2026-10-12 had 2,761,577 bytes, SHA-256 3d37b32e8ff782201ed8ee6677a01808d4f8c86eea70350a530dde7306cab7c9 and PDF CreationDate 2026-10-11T20:08:12-07:00. The reviewer's copy's text layer matched every extracted value. The licence does not allow archiving. |
| Comparison title CpGPT vs classical classifiers on EpiPanGI Dx cfDNA methylation (Xu and Feng 2026) Context-only references | An Evaluation of DMR-Informed Fine-Tuning of Tissue Array-Pretrained CpGPT for Gastrointestinal Cancer Classification Using cfDNA Targeted Methylation ctdnalit-20261011-cpgpt2026-source-preprint: Table 3; Methods; Results Version: medRxiv version 1, posted 2026-07-29 | not individually reviewed No individual claim review recorded Audit detailsField: Source artifact SHA-256: Hash scope: artifact_sha256 describes one retrieval (2026-10-11T21:49:59Z). medRxiv regenerates this PDF: a reviewer download on 2026-10-12 had 2,761,577 bytes, SHA-256 3d37b32e8ff782201ed8ee6677a01808d4f8c86eea70350a530dde7306cab7c9 and PDF CreationDate 2026-10-11T20:08:12-07:00. The reviewer's copy's text layer matched every extracted value. The licence does not allow archiving. |
Constraints3 values
| An Evaluation of DMR-Informed Fine-Tuning of Tissue Array-Pretrained CpGPT for Gastrointestinal Cancer Classification Using cfDNA Targeted Methylation ctdnalit-20261011-cpgpt2026-source-preprint: Table 3; Methods; Results Version: medRxiv version 1, posted 2026-07-29 | not individually reviewed No individual claim review recorded Audit detailsField: Source artifact SHA-256: Hash scope: artifact_sha256 describes one retrieval (2026-10-11T21:49:59Z). medRxiv regenerates this PDF: a reviewer download on 2026-10-12 had 2,761,577 bytes, SHA-256 3d37b32e8ff782201ed8ee6677a01808d4f8c86eea70350a530dde7306cab7c9 and PDF CreationDate 2026-10-11T20:08:12-07:00. The reviewer's copy's text layer matched every extracted value. The licence does not allow archiving. |
| Endpoint GI cancer vs non-cancer test AUROC and AUPRC (and validation AUROC), mean over three split-runs, for a fine-tuned methylation foundation model (CpGPT) and four classical classifiers on the same targeted bisulfite cfDNA features and partitions Context-only references | An Evaluation of DMR-Informed Fine-Tuning of Tissue Array-Pretrained CpGPT for Gastrointestinal Cancer Classification Using cfDNA Targeted Methylation ctdnalit-20261011-cpgpt2026-source-preprint: Table 3; Methods; Results Version: medRxiv version 1, posted 2026-07-29 | not individually reviewed No individual claim review recorded Audit detailsField: Source artifact SHA-256: Hash scope: artifact_sha256 describes one retrieval (2026-10-11T21:49:59Z). medRxiv regenerates this PDF: a reviewer download on 2026-10-12 had 2,761,577 bytes, SHA-256 3d37b32e8ff782201ed8ee6677a01808d4f8c86eea70350a530dde7306cab7c9 and PDF CreationDate 2026-10-11T20:08:12-07:00. The reviewer's copy's text layer matched every extracted value. The licence does not allow archiving. |
| Claimed field links:assessed_by:ctdnalit-20261011-cpgpt2026-protocol-dmr-informed-split-runs Context-only references | An Evaluation of DMR-Informed Fine-Tuning of Tissue Array-Pretrained CpGPT for Gastrointestinal Cancer Classification Using cfDNA Targeted Methylation ctdnalit-20261011-cpgpt2026-source-preprint: Table 3; Methods; Results Version: medRxiv version 1, posted 2026-07-29 | not individually reviewed No individual claim review recorded Audit detailsField: Source artifact SHA-256: Hash scope: artifact_sha256 describes one retrieval (2026-10-11T21:49:59Z). medRxiv regenerates this PDF: a reviewer download on 2026-10-12 had 2,761,577 bytes, SHA-256 3d37b32e8ff782201ed8ee6677a01808d4f8c86eea70350a530dde7306cab7c9 and PDF CreationDate 2026-10-11T20:08:12-07:00. The reviewer's copy's text layer matched every extracted value. The licence does not allow archiving. |
| Headline metric auroc Context-only references | An Evaluation of DMR-Informed Fine-Tuning of Tissue Array-Pretrained CpGPT for Gastrointestinal Cancer Classification Using cfDNA Targeted Methylation ctdnalit-20261011-cpgpt2026-source-preprint: Table 3; Methods; Results Version: medRxiv version 1, posted 2026-07-29 | not individually reviewed No individual claim review recorded Audit detailsField: Source artifact SHA-256: Hash scope: artifact_sha256 describes one retrieval (2026-10-11T21:49:59Z). medRxiv regenerates this PDF: a reviewer download on 2026-10-12 had 2,761,577 bytes, SHA-256 3d37b32e8ff782201ed8ee6677a01808d4f8c86eea70350a530dde7306cab7c9 and PDF CreationDate 2026-10-11T20:08:12-07:00. The reviewer's copy's text layer matched every extracted value. The licence does not allow archiving. |
Limitations4 values
| An Evaluation of DMR-Informed Fine-Tuning of Tissue Array-Pretrained CpGPT for Gastrointestinal Cancer Classification Using cfDNA Targeted Methylation ctdnalit-20261011-cpgpt2026-source-preprint: Table 3; Methods; Results Version: medRxiv version 1, posted 2026-07-29 | not individually reviewed No individual claim review recorded Audit detailsField: Source artifact SHA-256: Hash scope: artifact_sha256 describes one retrieval (2026-10-11T21:49:59Z). medRxiv regenerates this PDF: a reviewer download on 2026-10-12 had 2,761,577 bytes, SHA-256 3d37b32e8ff782201ed8ee6677a01808d4f8c86eea70350a530dde7306cab7c9 and PDF CreationDate 2026-10-11T20:08:12-07:00. The reviewer's copy's text layer matched every extracted value. The licence does not allow archiving. |
Pinned records23 values
| An Evaluation of DMR-Informed Fine-Tuning of Tissue Array-Pretrained CpGPT for Gastrointestinal Cancer Classification Using cfDNA Targeted Methylation ctdnalit-20261011-cpgpt2026-source-preprint: Table 3; Methods; Results Version: medRxiv version 1, posted 2026-07-29 | not individually reviewed No individual claim review recorded Audit detailsField: Source artifact SHA-256: Hash scope: artifact_sha256 describes one retrieval (2026-10-11T21:49:59Z). medRxiv regenerates this PDF: a reviewer download on 2026-10-12 had 2,761,577 bytes, SHA-256 3d37b32e8ff782201ed8ee6677a01808d4f8c86eea70350a530dde7306cab7c9 and PDF CreationDate 2026-10-11T20:08:12-07:00. The reviewer's copy's text layer matched every extracted value. The licence does not allow archiving. |
| Rationale Compares a methylation foundation model with conventional classifiers on the same plasma cfDNA methylation samples and splits. Proxy: AUROC without a declared specificity, very small test partitions, feature selection that saw the test samples, and an unreviewed preprint. Context-only references | An Evaluation of DMR-Informed Fine-Tuning of Tissue Array-Pretrained CpGPT for Gastrointestinal Cancer Classification Using cfDNA Targeted Methylation ctdnalit-20261011-cpgpt2026-source-preprint: Table 3; Methods; Results Version: medRxiv version 1, posted 2026-07-29 | not individually reviewed No individual claim review recorded Audit detailsField: Source artifact SHA-256: Hash scope: artifact_sha256 describes one retrieval (2026-10-11T21:49:59Z). medRxiv regenerates this PDF: a reviewer download on 2026-10-12 had 2,761,577 bytes, SHA-256 3d37b32e8ff782201ed8ee6677a01808d4f8c86eea70350a530dde7306cab7c9 and PDF CreationDate 2026-10-11T20:08:12-07:00. The reviewer's copy's text layer matched every extracted value. The licence does not allow archiving. |
Sources and history
Release 2026-11-21-c8bea25e5abb · Record review: source checked
1 source record and release history
- An Evaluation of DMR-Informed Fine-Tuning of Tissue Array-Pretrained CpGPT for Gastrointestinal Cancer Classification Using cfDNA Targeted Methylation · Original source · medRxiv version 1, posted 2026-07-29
Technical metadata and extraction receipts
Stable ID: use-case-mapping-ctdna-lit-20261011-cpgpt2026-split-runs
- field
- links:assessed_by:ctdnalit-20261011-cpgpt2026-protocol-dmr-informed-split-runs
- value
- ctdnalit-20261011-cpgpt2026-protocol-dmr-informed-split-runs
- relevance
- proxy
- endpoint
- GI cancer vs non-cancer test AUROC and AUPRC (and validation AUROC), mean over three split-runs, for a fine-tuned methylation foundation model (CpGPT) and four classical classifiers on the same targeted bisulfite cfDNA features and partitions
- rationale
- Compares a methylation foundation model with conventional classifiers on the same plasma cfDNA methylation samples and splits. Proxy: AUROC without a declared specificity, very small test partitions, feature selection that saw the test samples, and an unreviewed preprint.
- constraints
- Inspect every linked evaluation's source locator and preserved limitations before citing a result.; Do not combine with other protocols' results; assay, cohort and features differ.; Treat differences between models as descriptive: three split-runs of about 30 test samples each.
- limitations
- The 20,499 candidate CpGs were selected using cancer-versus-control labels of all 300 samples, including the later test partitions, so test performance is not independent of feature selection (stated by the authors).; Each test partition is about 10% of 300 samples (roughly 30, with few controls); means and 95% CIs are over three split-runs only, bounded to 0-1.; Clinically diagnosed gastrointestinal cancers against 46 non-cancer controls; not a screening population.; Preprint, not peer reviewed; the authors call the model comparisons descriptive.
- citation locators
- source id: ctdnalit-20261011-cpgpt2026-source-preprint; locator: Table 3; Methods; Results
- source locator
- ctdnalit-20261011-cpgpt2026-source-preprint: Table 3; Methods; Results
- revision
- 1
- reason
- Add the CpGPT versus classical-classifier comparison on cfDNA targeted methylation (Xu and Feng 2026, medRxiv), retrieved 2026-10-11. Independently reviewed in docs/reviews/use-cases/plasma-ctdna-use-cases-2026-10-12.md.
- comparison group
- cpgpt2026-gse149438
- comparison title
- CpGPT vs classical classifiers on EpiPanGI Dx cfDNA methylation (Xu and Feng 2026)
- headline metric
- auroc
- review
- method: independent-cell-check; scope-review; 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: 3d37b32e8ff782201ed8ee6677a01808d4f8c86eea70350a530dde7306cab7c9; retrieval url: https://www.medrxiv.org/content/10.64898/2026.07.26.26358979v1.full.pdf; note: The pinned bytes could not be re-verified: medRxiv regenerates this PDF, and a download on 2026-10-12 had a different SHA-256 (see the source's hash_scope). Compared all 15 Table 3 values and intervals, the asserted sentences and the footnote with the layout text of that copy, read for this review (the extractor's script was not run); all match.
- reviewed evaluations
- ctdnalit-20261011-cpgpt2026-eval-cpgpt; ctdnalit-20261011-cpgpt2026-eval-elastic-net; ctdnalit-20261011-cpgpt2026-eval-gbm; ctdnalit-20261011-cpgpt2026-eval-random-forest; ctdnalit-20261011-cpgpt2026-eval-svm-radial
- pins
- record id: ctdnalit-20261011-cpgpt2026-eval-cpgpt; fields: status; links; attributes.origin; attributes.comparison; sha256: 78f9ad609a26d5f54d2713b0fc7cbea37511250f9c01248f57935b7dbeba7df9; record id: ctdnalit-20261011-cpgpt2026-eval-elastic-net; fields: status; links; attributes.origin; attributes.comparison; sha256: a18c7c946da60f42bb28b95e303c0fdf63f35d0097fae1d1938496199cbdc501; record id: ctdnalit-20261011-cpgpt2026-eval-gbm; fields: status; links; attributes.origin; attributes.comparison; sha256: 544558b9caedc6b1bd16b7accc0f6cdda15567ab0c4c58220514dbd1eb7ac145; record id: ctdnalit-20261011-cpgpt2026-eval-random-forest; fields: status; links; attributes.origin; attributes.comparison; sha256: b4633bec86dd51af8a795bd5522057e4f6022802d34682785824ccb54cfda13c; record id: ctdnalit-20261011-cpgpt2026-eval-svm-radial; fields: status; links; attributes.origin; attributes.comparison; sha256: 4da849c20f8388275f808441a542f5c14315b841e45e87081e24b403310bdb1f; record id: ctdnalit-20261011-cpgpt2026-protocol-dmr-informed-split-runs; fields: status; name; description; facets; source_ids; links; attributes; sha256: 507216e93a0c73a430847f1c118c5d24e9915b7b702854a7bce6886600a2ee45; record id: ctdnalit-20261011-cpgpt2026-result-cpgpt-test-auprc; fields: status; attributes.metric; attributes.metric_qualifier; attributes.printed_value; attributes.numeric_value; sha256: 09e19d586589be978e57efe26f8a717bcfbb78e4956192d33ef3bd0d7abb1ca2; record id: ctdnalit-20261011-cpgpt2026-result-cpgpt-test-auroc; fields: status; attributes.metric; attributes.metric_qualifier; attributes.printed_value; attributes.numeric_value; sha256: 58d30d63b528e2c0e5c71c70d10d676fa63085370a120d91fc833d3c5d685b81; record id: ctdnalit-20261011-cpgpt2026-result-cpgpt-validation-auroc; fields: status; attributes.metric; attributes.metric_qualifier; attributes.printed_value; attributes.numeric_value; sha256: 7cfc406fe2e53509df5e4fce529a7bacc071e9172cf16c08b4f8dd4545611c64; record id: ctdnalit-20261011-cpgpt2026-result-elastic-net-test-auprc; fields: status; attributes.metric; attributes.metric_qualifier; attributes.printed_value; attributes.numeric_value; sha256: f85303d508279950bccbac0c5c775d5496c6b802ed66d3fb0457949852032a1b; record id: ctdnalit-20261011-cpgpt2026-result-elastic-net-test-auroc; fields: status; attributes.metric; attributes.metric_qualifier; attributes.printed_value; attributes.numeric_value; sha256: 56009e5c47c2b4d9952418eaf2e2035963bde294693dbf5b5ac4798ad2640b7b; record id: ctdnalit-20261011-cpgpt2026-result-elastic-net-validation-auroc; fields: status; attributes.metric; attributes.metric_qualifier; attributes.printed_value; attributes.numeric_value; sha256: 11b9722a4beb923fc9eb06958d8cbddf2c9887c720776eecfbb41681ff6b3eb1; record id: ctdnalit-20261011-cpgpt2026-result-gbm-test-auprc; fields: status; attributes.metric; attributes.metric_qualifier; attributes.printed_value; attributes.numeric_value; sha256: 63469b42bdc30771e94e5c128784f63593a546e809b100290f3faaea77c161ac; record id: ctdnalit-20261011-cpgpt2026-result-gbm-test-auroc; fields: status; attributes.metric; attributes.metric_qualifier; attributes.printed_value; attributes.numeric_value; sha256: 50848c670daaa08ba691c2509144669398c591247f9e16039e421d242b2f7239; record id: ctdnalit-20261011-cpgpt2026-result-gbm-validation-auroc; fields: status; attributes.metric; attributes.metric_qualifier; attributes.printed_value; attributes.numeric_value; sha256: 7db097c7ffd15372bf6031861b7128fd443532c9c47bc7a185ffa78ef42db913; record id: ctdnalit-20261011-cpgpt2026-result-random-forest-test-auprc; fields: status; attributes.metric; attributes.metric_qualifier; attributes.printed_value; attributes.numeric_value; sha256: f1b44c1ddb5c1d3b3dcaee6801fbeeddcd5aad5e0a87b3f5fe8689805524fb47; record id: ctdnalit-20261011-cpgpt2026-result-random-forest-test-auroc; fields: status; attributes.metric; attributes.metric_qualifier; attributes.printed_value; attributes.numeric_value; sha256: fae2c89bc394f30e6659d9a58b0af8afb18be2a9b09a6a4b785176a295ea4c31; record id: ctdnalit-20261011-cpgpt2026-result-random-forest-validation-auroc; fields: status; attributes.metric; attributes.metric_qualifier; attributes.printed_value; attributes.numeric_value; sha256: ce4baa67d2df820bd69a3443f4f6f9b4a54eee7b3abae93ab6688e884328d917; record id: ctdnalit-20261011-cpgpt2026-result-svm-radial-test-auprc; fields: status; attributes.metric; attributes.metric_qualifier; attributes.printed_value; attributes.numeric_value; sha256: d6ba63fe6b8f0d2a999c0d6025d06b8ea5cda10998cf5fc9e8eb4e050749196b; record id: ctdnalit-20261011-cpgpt2026-result-svm-radial-test-auroc; fields: status; attributes.metric; attributes.metric_qualifier; attributes.printed_value; attributes.numeric_value; sha256: 8b8976de2e5c426b4c170f73638a642eb248587ab0f175ad7e61b87b9c1422cf; record id: ctdnalit-20261011-cpgpt2026-result-svm-radial-validation-auroc; fields: status; attributes.metric; attributes.metric_qualifier; attributes.printed_value; attributes.numeric_value; sha256: 43c7e0c82c50c30f5bb5ed7df99091ce41251f90a6217ae842bf8dc0c1913710; record id: ctdnalit-20261011-cpgpt2026-source-preprint; fields: status; attributes.artifact_sha256; sha256: 248745d91ebac2cb3f581f75cb5473254aa11afb26a82d4b53c74f72dbc67ed5; record id: use-case-plasma-ctdna-methylation; fields: status; name; description; facets; attributes.question; attributes.decision; attributes.inputs; attributes.output; attributes.exclusions; attributes.setting; attributes.clinical_scope; attributes.intended_users; sha256: 4d9542463e1400d5d9701c7f52859a23e4c78804cbda1477663445f945e22429