Identification and Validation of Glycosylation‑Related Genes in Ischemic Stroke Based on Bioinformatics and Machine Learning

Abstract

Ischemic stroke (IS) constitutes a severe neurological disorder with restricted treatment alternatives. Recent investigations have disclosed that glycosylation is closely associated with the occurrence and outcome of IS. Nevertheless, data on the transcriptomic dynamics of glycosylation in IS are lacking. The objective of this study was to undertake a comprehensive exploration of glycosylation-related genes (GRGs) in IS via bioinformatics and to assess their immune characteristics. In this study, through the intersection of genes from weighted gene co-expression network analysis, GRGs from five glycosylation pathways, and DEGs from differential expression analysis, 20 candidate GRGs were identified. Subsequently, through LASSO, Random Forest, and SVM-RFE, 3 hub GRGs (F5, PPP6C, and UBE2J1) were identified. Additional, a gene diagnostic model linked to glycosylation was developed and validated. The findings indicated that the diagnostic model could effectively distinguish between IS patients and healthy individuals in the training, validation, and merging datasets, indicating clinical relevance. Subsequently, by employing unsupervised clustering analysis, IS patients were classified into three clusters, and significant disparities were witnessed in immune cell infiltration among distinct clusters. In summary, this study successfully identified hub GRGs in IS and investigated the roles of these hub genes in the immune microenvironment, indicating potential clinical applications for IS.

Keywords: Glycosylation; Hub gene; Immune cell infiltration; Ischemic Stroke; Machine learning.