AI Science Tracker
EN

Learned transcriptomic selection identifies a pediatric-sepsis blood-marker pattern

Zhejiang University; Shaoxing People’s Hospital; Columbia University · 28 Apr 2023 · biology

Discovery

Learned transcriptomic selection identifies a pediatric-sepsis blood-marker pattern

Source check: Source checked. Not independent scientific confirmation.

Finding: Three learned feature-selection approaches converged on CYSTM1, MMP8 and CD177 in public pediatric blood-expression data. The source reports higher expression and case-control discrimination in a further dataset. This is a retrospective molecular association, including a less-studied CYSTM1 candidate, without established diagnostic specificity, clinical utility or causal treatment relevance.

Technology: Random forest, SVM recursive feature elimination and LASSO logistic feature selection

Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.

AI contribution
The three learned selectors prioritized a shared measured-expression marker set after broad filtering, enabling the represented candidate-pattern and subsequent source-described case-control comparison. Immune deconvolution or network diagrams alone are not the admission basis.
Test and result
The analysis combined GSE26378 and GSE26440, with primary series metadata listing 180 septic-shock and 53 normal-control blood-expression profiles in total. Differential/module filtering preceded RF, SVM-RFE and tenfold-tuned LASSO selection. Their intersection identified CYSTM1, MMP8 and CD177. The paper reports increased expression and discrimination in GSE13904, a separately indexed pediatric inflammatory-spectrum dataset.
Limitation / still unknown
Distinct series/sample identifiers do not prove independent people. GSE13904 includes inflammatory-response, sepsis and septic-shock samples; the exact analytic subset and repeated-person separation were not independently reconstructed. Feature filtering used clinical labels before selection, and healthy-control contrasts do not establish real-world diagnostic specificity. Source prose twice refers to rheumatoid arthritis despite the sepsis analysis; no RA result is adopted. MMP8/CD177 links to sepsis were already known, and CYSTM1 remains a candidate association without functional or therapeutic validation. Reported ROC/decision curves do not demonstrate clinical benefit.
Evidence sources and annotation context

Evidence context (descriptions follow the cited sources; test type is a catalog classification):

Evidence sources: https://pmc.ncbi.nlm.nih.gov/articles/PMC10533616/ · https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE26378 · https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE26440 · https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE13904

Source checked · editorial details

The source page was checked against this entry. Independent confirmation is not assessed by this label.

Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.

Review date: 21 Nov 2026. Review scope: Read original computational cohort/filtering/RF/SVM-RFE/LASSO and ROC Methods, candidate/validation-expression Results and stated limits; separately read primary GEO series-level cohort metadata and exact GSM intersections. Patient-level independence and full validation subset were not reconstructed. No operational biological procedures or causal treatment claim. Exact original article front DOI/title and fresh PubMed primary relationship metadata checked..

No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.

Editorial assessment: Impact 1/5 · Validation 2/5 · Novelty 2/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.

Open in the full catalog for filters, charts and related discoveries. Browse biology discoveries.

Related catalog entries

Links use documented relationships where available, plus shared catalog context. A shared topic or model family does not establish a scientific dependency.