AI Science Tracker
EN

Learned metabolite selection supports a thrombosis classification panel in coronary-heart-disease patients

Xi’an Jiaotong University; Honghui Hospital; Federal University Dutsin-Ma · 28 Mar 2026 · medicine

Discovery

Learned metabolite selection supports a thrombosis classification panel in coronary-heart-disease patients

Source check: Source checked. Not independent scientific confirmation.

Finding: LASSO and random-forest selection identified a six-metabolite panel associated with Doppler-defined post-traumatic deep-vein thrombosis among patients with coronary heart disease. The source reports evaluation of the resulting classifier in a separate same-center clinical cohort. These are retrospective biomarker associations and diagnostic discrimination; the study does not establish that the metabolites cause thrombosis or that predicted drug candidates are effective.

Technology: LASSO regression and random-forest feature selection; logistic classification

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

AI contribution
Learned sparse regression and RF feature ranking selected the six-metabolite subset from measured profiles, enabling construction and separate-cohort evaluation of the represented clinical classification panel.
Test and result
The source describes 218 discovery participants, comprising 93 thrombosis-plus-coronary-disease cases and 125 coronary-disease controls, and a separate 86-person validation cohort with 39 cases and 47 controls. A 2000-tree RF and LASSO selected six overlapping metabolite features, used in logistic diagnostic models. The original Methods describe clinical Doppler reference assessment, model construction and evaluation in the separate cohort.
Limitation / still unknown
Both cohorts came from one hospital and underwent different analytical batches. Differential screening preceded internal train/test modeling; some preprocessing used dataset-wide information, so internal split results are not a wholly untouched pipeline evaluation. The paper describes different internal split ratios for feature-selection and final-model stages. Blood sampling and Doppler confirmation were not simultaneous in every patient, and causal pathway interpretation is not established. Metabolite/protein totals vary across source sections; those totals and simulated therapeutic candidates are not represented as verified findings.
Evidence sources and annotation context

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

Evidence sources: https://link.springer.com/article/10.1007/s11306-026-02419-3

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 actual clinical recruitment/reference timing, discovery/validation groups, preprocessing/feature-selection/model Methods, six-feature intersection and diagnostic-model Results. Causal sphingolipid and drug-screen claims are excluded from the represented finding. Original canonical online March 28 distinguished from April issue. 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 2/5 · 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 medicine 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.