Identification of epileptic hippocampal sclerosis related genes through bulk and single-nucleus RNA sequencing datasets.

BACKGROUND: Hippocampal sclerosis is one of the most common pathological alterations in drug-refractory epilepsy. However, the exact causes and molecular mechanisms of its formation remain unknown. METHODS: We constructed a hippocampal sclerosis score (HSS) based on the collective expression of selected hub genes harvested from multiple transcriptomic datasets. We also evaluated immune infiltration and related hallmark signaling pathways to verify the connection between immune cells and hippocampal sclerosis pathogenesis. RESULTS: Based on three machine learning models and differential expressed genes (DEGs) from multiple RNA-seq datasets, four intersecting genes were identified (PSD4, CFAP47, TMEM156 and P2RY13). We found significant infiltration of immune cells via the CIBERSORT algorithm, particularly in myeloid cell features. Diagnostic nomogram was established to predict the incidence of hippocampal sclerosis. Single-nucleus RNA-seq (snRNA-seq) data elucidated the distinct signaling pathways of microglia based on HSS. CONCLUSION: We excavated an HSS model from multiple bulk RNA and snRNA-seq datasets. This model was used to evaluate the likelihood of developing hippocampal sclerosis and reveal potential factors such as signaling changes and immune infiltrations during hippocampal sclerosis formation.