Integrative identification of immune-related key genes in atrial fibrillation using weighted gene coexpression network analysis and machine learning

• DOI: 10.3389/fcvm.2022.922523
• Publication Date: 2022-07-27T07:13:55Z
• Authors: Peng-Fei Zheng; Lu-Zhu Chen; Peng Liu; Zheng-Yu Liu; Hong Wei Pan
• Keywords: Lasso

The immune system significantly participates in the pathologic process of atrial fibrillation (AF). However, molecular mechanisms underlying this participation are not completely explained. current research aimed to identify critical genes and cells that participate AF.CIBERSORT was utilized reveal cell infiltration pattern AF patients. Meanwhile, weighted gene coexpression network analysis (WGCNA) meaningful modules were correlated with AF. characteristic identified by least absolute shrinkage selection operator (LASSO) logistic regression support vector machine recursive feature elimination (SVM-RFE) algorithm.In comparison sinus rhythm (SR) individuals, we observed fewer activated mast regulatory T (Tregs), as well more gamma delta cells, resting M2 macrophages, infiltrated Three significant (pink, red, magenta) be associated Gene enrichment showed all 717 immunity- or inflammation-related pathways biological processes. Four hub (GALNT16, HTR2B, BEX2, RAB8A) revealed SVM-RFE algorithm LASSO regression. qRT-PCR results suggested compared SR subjects, patients exhibited reduced BEX2 GALNT16 expression, dramatically elevated HTR2B expression. AUC measurement diagnostic efficiency training set 0.836, 0.883, 0.893, respectively, 0.858, 0.861, 0.915, validation set.Three novel genes, GALNT16, WGCNA combined learning, which provides potential new therapeutic targets for early diagnosis prevention