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.