Identification and validation of key genes as diagnostic biomarkers for epilepsy using bioinformatics and machine...

Epilepsy is a common neurological disorder with a complex molecular pathogenesis. Approximately one-third of patients experience drug-resistant seizures, which severely impairs their quality of life. This study employed an integrated bioinformatics and machine learning approach to identify key genes associated with epilepsy. Epilepsy-related datasets from the GEO database were analyzed using differential expression analysis and weighted gene co-expression network analysis (WGCNA) for the initial identification of candidate genes. These candidates were subsequently refined using machine learning algorithms, including least absolute shrinkage and selection operator (LASSO), random forest (RF) and support vector machine (SVM), to filter key genes. TYMP and HES4 demonstrated robust diagnostic value for epilepsy (AUC: 0.770 for TYMP and 0.769 for HES4). Functional enrichment analysis indicated their involvement in key signaling pathways and a close association with immune cell infiltration, highlighting their role in modulating the immune microenvironment. The significant upregulation of TYMP and HES4 was validated in both epileptic mouse models and epilepsy cell lines using RT-qPCR and Western blot. Moreover, screening of a drug-gene interaction database identified Calpain inhibitor I as a potential therapeutic agent targeting TYMP. Overall, these findings provide novel insights into the pathogenesis of epilepsy and suggest promising targets for future research and therapeutic development.