Researchers investigating coronary artery disease have identified two distinct molecular subtypes associated with adipogenesis, according to a study published in the Journal of Cardiothoracic Surgery. The investigative team applied machine learning algorithms and bioinformatics analysis to uncover how adipogenesis-related gene expression correlates with coronary disease mechanisms. Adipogenesis functions as an intricate biological network involving diverse transcription factors and signaling pathways, which accentuates complex interactions in coronary artery disease including inflammation, insulin resistance, and microvascular dysfunction.
To establish these molecular classifications, the scientists analyzed transcriptional data obtained from the GSE20680 dataset. From this bioinformatics screening of adipogenesis-associated genetic factors, they identified 14 differentially expressed adipogenesis-related genes. Machine learning models were subsequently applied to evaluate these candidates, ultimately determining that two specific hub genes, CD36 and COX7B, served as effective predictive markers for coronary artery disease.
Following the selection of CD36 and COX7B, the investigators devised a clinical nomogram incorporating both regulatory genes to predict disease risk. The predictive accuracy of this scoring tool was evaluated and validated through decision curve analysis. By applying a consensus clustering algorithm to the genetic profiles, the researchers formally classified patients with coronary artery disease into two distinct biological subtypes, allowing further exploration of how these profiles interface with immune responses.
The bioinformatic findings were subsequently validated using human clinical specimens. Enzyme-linked immunosorbent assays performed on patient serum samples revealed elevated circulating levels of both CD36 and COX7B in individuals diagnosed with coronary artery disease. These measurements demonstrated a consistent expression trend matching the elevated patterns originally observed in the transcriptomic data.
To examine the functional consequences of these candidate genes in a living organism, the research team constructed an experimental zebrafish model. Through this animal model, the researchers explored how the overexpression of CD36 and COX7B alters lipid metabolism and vascular development. These experimental observations provided biological context linking the machine learning predictions to tangible cardiovascular alterations.
The study authors concluded that adipogenesis-related processes play a defined role in the prevalence and presentation of coronary artery disease. Identifying distinct adipogenesis subtypes provides a biological framework that holds potential for informing future personalized management and treatment strategies. Furthermore, the characterization of CD36 and COX7B as key hub genes suggests these regulators may serve as promising molecular targets for developing prospective cardiovascular therapeutics.
Subject of Research: Machine learning identification of adipogenesis-related coronary artery disease subtypes
Article Title: Identification and validation of the adipogenesis-related molecular subtypes of coronary artery disease by machine learning
Article References: Gu, J., Lu, T., Wu, Y., Shi, H., Chen, X., & Zhang, S. (2026). Identification and validation of the adipogenesis-related molecular subtypes of coronary artery disease by machine learning. Journal of Cardiothoracic Surgery. https://doi.org/10.1186/s13019-026-04848-6
Image Credits: AI Generated
DOI: 10.1186/s13019-026-04848-6
Keywords: coronary artery disease, adipogenesis, machine learning, CD36, COX7B, molecular subtypes, consensus clustering, bioinformatics, nomogram, decision curve analysis, ELISA, serum biomarkers, zebrafish model, lipid metabolism, microvascular dysfunction, insulin resistance, cardiovascular disease
News Source: Teresa Odom. (November 19, 2026). Machine Learning Identifies Two Adipogenesis Subtypes in Coronary Artery Disease. Scienmag.



