Potential Diagnostic Biomarkers in Heart Failure: Suppressed Immune-Associated Genes Identified by Bioinformatic Analysis and Machine Learning.

Abstract

Heart failure (HF) threatens tens of millions of people's health worldwide, which is the terminal stage in the development of majority cardiovascular diseases. Recently, an increasing number of studies have demonstrated that bioinformatics and machine learning (ML) algorithms can offer new insights into the diagnosing and treating HF. To further discover HF diagnostic genes, we utilized least absolute shrinkage and selection operator (LASSO) and Support Vector Machine (SVM) to identify novel immune-related genes. The HF dataset was obtained from the gene expression omnibus (GEO) database and three candidate genes (LCN6, MUC4, and TNFRSF13C) were finally screened. In addition, the myocardial infarction (MI) modeling experiments on mice were performed to validate the expression of LCN6, MUC4, and TNFRSF13C on experimental HF mice. Altogether, these three candidate genes are promising targets for the prediction of HF with immunological perspective.