Abstract
Schizophrenia (SCZ), which is characterized by debilitating neuropsychiatric disorders with significant cognitive impairment, remains an etiological and therapeutic challenge. Using transcriptomic profile analysis, disease-related biomarkers linked with SCZ have been identified, and clinical outcomes can also be predicted. This study aimed to discover diagnostic hub genes and investigate their possible involvement in SCZ immunopathology. The Gene Expression Omnibus (GEO) database was utilized to get SCZ Gene expression data. Differentially expressed genes (DEGs) were identified and enriched by Gene Ontology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG), and disease ontology (DO) analysis. The related gene modules were then examined using integrated weighted gene co-expression network analysis. Single-sample gene set enrichment (GSEA) was exploited to detect immune infiltration. SVM-REF, random forest, and least absolute shrinkage and selection operator (LASSO) algorithms were used to identify hub genes. A diagnostic model of nomogram was constructed for SCZ prediction based on the hub genes. The clinical utility of nomogram prediction was evaluated, and the diagnostic utility of hub genes was validated. mRNA levels of the candidate genes in SCZ rat model were determined. Finally, 24 DEGs were discovered, the majority of which were enriched in biological pathways and activities. Four hub genes (NEUROD6, NMU, PVALB, and NECAB1) were identified. A difference in immune infiltration was identified between SCZ and normal groups, and immune cells were shown to potentially interact with hub genes. The hub gene model for the two datasets was verified, showing good discrimination of the nomogram. Calibration curves demonstrated valid concordance between predicted and practical probabilities, and the nomogram was verified to be clinically useful. According to our research, NEUROD6, NMU, PVALB, and NECAB1 are prospective biomarkers in SCZ and that a reliable nomogram based on hub genes could be helpful for SCZ risk prediction.