Identification of endoplasmic reticulum stress genes in human stroke based on bioinformatics and machine learning

Nan Jiang,Chuying Wang,Bingqing Xie,Huangfan Xie,Anguo Wu,Xi Kong,Long Gu,Yong Jiang,Jianhua Peng
DOI: https://doi.org/10.1016/j.nbd.2024.106583
Abstract:After ischemic stroke (IS), secondary injury is intimately linked to endoplasmic reticulum (ER) stress and body-brain crosstalk. Nonetheless, the underlying mechanism systemic immune disorder mediated ER stress in human IS remains unknown. In this study, 32 candidate ER stress-related genes (ERSRGs) were identified by overlapping MSigDB ER stress pathway genes and DEGs. Three Key ERSRGs (ATF6, DDIT3 and ERP29) were identified using LASSO, random forest, and SVM-RFE. IS patients with different ERSRGs profile were clustered into two groups using consensus clustering and the difference between 2 group was further explored by GSVA. Through immune cell infiltration deconvolution analysis, and middle cerebral artery occlusion (MCAO) mouse scRNA analysis, we found that the expression of 3 key ERSRGs were closely related with peripheral macrophage cell ER stress in IS and this was further confirmed by RT-qPCR experiment. These ERS genes might be helpful to further accurately regulate the central nervous system and systemic immune response through ER stress and have potential application value in clinical practice in IS.
What problem does this paper attempt to address?