Exploring Key Genes with Diagnostic Value for Nonalcoholic Steatohepatitis Based on Bioinformatics Analysis

We aimed to screen specific genes in liver tissue samples of patients with nonalcoholic steatohepatitis (NASH) with clinical diagnostic value based on bioinformatics analysis. The datasets of liver tissue samples from healthy individuals and NASH patients were retrieved for consistency cluster analysis to obtain the NASH sample typing, followed by verification of the diagnostic value of sample genotyping-specific genes. All samples were subjected to logistic regression analysis, followed by the establishment of the risk model, and then, the diagnostic value was determined by receiver operating characteristic curve analysis. NASH samples could be divided into cluster 1, cluster 2, and cluster 3, which could predict the nonalcoholic fatty liver disease activity score of patients. A total of 162 sample genotyping-specific genes were extracted from patient clinical parameters, and the top 20 core genes in the protein interaction network were obtained for logistic regression analysis. Five sample genotyping-specific genes (WD repeat and HMG-box DNA-binding protein 1 [WDHD1], GINS complex subunit 2 [GINS2], replication factor C subunit 3 (RFC3), secreted phosphoprotein 1 [SPP1], and spleen tyrosine kinase [SYK]) were extracted to construct the risk models with high diagnostic value in NASH. Compared with the low-risk group, the high-risk group of the model showed increased lipoproduction and decreased lipolysis and lipid β oxidation. The risk models based on WDHD1, GINS2, RFC3, SPP1, and SYK have high diagnostic value in NASH, and this risk model is closely related to lipid metabolism pathways.