Identification and validation of INHBE and P4HA1 as hub genes in non-alcoholic fatty liver disease

Non-alcoholic fatty liver disease (NAFLD) is currently the most prevalent type of liver disease and a worldwide disease threatening human health. This study aims to identify the novel diagnostic biomarkers of NAFLD by comprehensive bioinformatics and machine learning, and to validate our results in hepatocyte and animal models. We used Gene Expression Omnibus (GEO) databases on NAFLD patients for differential gene expression analyses. Intersections were taken with genes from the key modules of WGCNA and differentially expressed genes (DEGs). Machine learning algorithms like LASSO regression analysis, SVM-RFE, and RandomForest were used to screen hub genes. In addition, a nomogram model and calibration curves were built in order to forecast the probability of NAFLD occurrence. Then, the relationship between hub genes and immune cells was verified using Spearman analysis. Finally, we further verified the expression of key genes by constructing a steatosis hepatocyte model and animal model. Key genes (INHBE and P4HA1) were identified by comprehensive bioinformatics analysis and machine learning. INHBE and P4HA1 were up-regulated and down-regulated in the steatosis hepatocyte model, respectively. Animal experiments also showed that INHBE was up-regulated in the liver of mice fed with high fat diet (HFD). INHBE and P4HA1 are the hub genes of NAFLD. Our findings may contribute to a greater understanding of the occurrence and development of NAFLD and provide potential biomarkers and possible therapeutic targets for future clinical diagnosis and treatment.