Abstract
BACKGROUND: Non-alcoholic fatty liver disease (NAFLD) is a manifestation of metabolic syndrome in the liver with varying severity. Heterogeneity in terms of molecules and immune cell infiltration drives NAFLD from one stage to the next. However, a precise molecular classification of NAFLD is still lacking, and the effects of complex clinical phenotypes on the efficacy of drugs are usually ignored. METHODS: We introduced multiple omics data to differentiate NAFLD subtypes via consensus clustering, and a weighted gene co-expression network analysis was used to identify eight co-expression modules. Further, eigengenes of eight modules were analyzed with regard to Gene Ontology enrichment and Kyoto Encyclopedia of Genes and Genomes pathways. Furthermore, the infiltration rates of 22 immune cell types were calculated with CIBERSORT and the ESTIMATE algorithm. RESULTS: In total, 111 NAFLD patients from three independent GEO datasets were divided into four molecular subtypes, and the corresponding clinical features and immune cell infiltration traits were determined. Based on high gene expression correlations, four molecular subtypes were further divided into eight co-expression modules. We also demonstrated a significant correlation between gene modules and clinical phenotypes. Moreover, we integrated phenotypic, immunologic, and genetic data to assess the potential for progression of different molecular subtypes. Furthermore, the efficacy of drugs against various NAFLD molecular subtypes was discussed to aid in individualized therapy. CONCLUSION: Overall, this study could provide new insights into the underlying pathogenesis of and drug targets for NAFLD.