Construction of a diagnostic prediction model for ischemic stroke using lactylation-related genes

Ischemic stroke (IS) represents the leading global cause of acquired neurological disability and vascular-related mortality. However, diagnostic challenges persist in cases with atypical presentations. Lactylation modification exerts critical regulatory roles in disease pathogenesis and progression, and thus positioning as a potential diagnostic biomarker. We utilized weighted gene co-expression network analysis (WGCNA), gene ontology (GO)and Kyoto Encyclopedia of Genes and Genomes (KEGG), immune infiltration assessment, consensus clustering (via ConsensusClusterPlus), and multiple machine learning algorithms-including random forest (RF), support vector machine (SVM), neural network (NM), and generalized linear models (GLMs)-along with real-time-quantitative polymerase chain reaction (RT-qPCR) and western blot validation, to analyze gene expression omnibus (GEO) datasets. Our findings indicate that immune infiltration may play an important role in IS, with neutrophils and T cell receptor signaling pathway emerging as the most important immune cells and signaling pathway, respectively. Six hub genes, namely SLC2A3, NDUFB11, GTPBP3, SLC16A3, PUS1, and GRN, were identified and verified through RT-qPCR and the western blot. Surprisingly, the area under the curve (AUC) of the prediction model reached 0.968, with a 95% confidence interval ranging from 0.928 to 1. Extensive validation using multiple external GEO datasets confirmed the accuracy of the prediction model in five independent datasets. Furthermore, we observed that different concentrations of lactate could further suppress the proliferation of nerve cells following oxygen-glucose deprivation/reperfusion (OGD/R). This study provides a new diagnostic strategy for the early diagnosis of IS through the established diagnostic prediction model.