Canonical text (the exact bytes fingerprinted)

TRUVACE RECORD VERSION
record: TRV-2026-1966
version: 1
kind: certified
reason: Certified into the record
timestamp: 2026-10-20T01:10:54.180786Z
status: published
lens: g_space
sector: health
headline: Genome-Wide DNA methylation profiling of tissue-plasma pairs identifies Novel cfDNA biomarkers for the diagnosis of gastric cancer
dek: Screening for gastric cancer (GC) is critical for improving patient survival, yet current clinical methods suffer from invasiveness and limited sensitivity. The aim of this study is to develop a machine learning model for the non-invasive diagnosis of GC using the whole-genome methylation profile of circulating free DNA (cfDNA) in plasma. Plasma cfDNA from 61 patients with GC and 29 healthy controls. First, whole‑genome bisulfite sequencing (WGBS) was performed on 29 paired GC and adjacent normal tissues to iden…
gain_title: A random forest model using eight plasma cfDNA methylation markers distinguished gastric cancer patients from healthy controls with an AUC of 0.938 on an independent test set, offering a non-invasive screening route.
problem_title: (none)
trace_subject: (none)
gain_reading: A random forest model using eight plasma cfDNA methylation markers distinguished gastric cancer patients from healthy controls with an AUC of 0.938 on an independent test set, offering a non-invasive screening route.
gain_evidence: achieving AUC values of 0.978 and 0.938 in the training and independent test sets
problem_reading: (none)
problem_evidence: (none)
quick_read: Researchers combined whole-genome bisulfite sequencing of 29 paired gastric cancer and adjacent normal tissues with plasma cfDNA methylation data from 61 patients and 29 healthy controls. They found 32 DMRs that changed consistently in tissue and plasma, narrowed them to eight markers, and trained four machine learning classifiers.

The work matters because current gastric cancer screening is invasive and has limited sensitivity, so a blood-based test could widen access to early detection. The evidence comes from a small cohort and a test set of limited size, so clinical usefulness is still unproven and the specificity leaves room for improvement.
limitation: The cohort is small (61 patients with gastric cancer and 29 healthy controls), and specificity on the test set was 75.00%, so the model is described only as promising rather than validated for screening.
tag: Evidence-backed gain
key_points: Random forest outperformed support vector machine, K-nearest neighbor and Naive Bayes models. | Test-set accuracy was 88.46%, sensitivity 94.44% and specificity 75.00%. | After age adjustment, the test-set AUC was 0.924. | The core markers lie in regions not covered by the 450K and EPIC v2 methylation arrays. | Markers were linked to phagocytosis and axon guidance pathways and to immune cell infiltration.
rundown: The study first used whole-genome bisulfite sequencing on paired tumor and adjacent normal tissue to map differentially methylated regions. These were matched against plasma cfDNA data, producing 32 stable regions that shifted in the same direction in both sample types. Tissue showed global hypomethylation and CpG island hypermethylation, consistent with the plasma patterns.

LASSO and random forest feature selection reduced the panel to eight markers, annotated to genes including ATP8A2. Functional analysis tied them to phagocytosis and axon guidance and to immune cell infiltration. The authors say the regions are missed by commonly used arrays, which they present as an advantage of whole-genome sequencing for discovering new biomarkers.
sources:
- peer_reviewed | PLOS Genetics | https://doi.org/10.1371/journal.pgen.1012322 | 2026-10-15
prev: 0000000000000000000000000000000000000000000000000000000000000000

sha256

ab7ccf27e3e251f29123c8fb14d4061ec3dd25009a096a8c2d1c94af5f415348

previous

0000000000000000000000000000000000000000000000000000000000000000