Alcohol-associated liver disease is one of the most prevalent diseases around the world, with 10-20% of patients developing progressive liver disease. To identify the complex and correlated nature of metabolic and microbial data types in progressive liver disease, we performed an integrated analysis of the fecal and serum metabolomes with the gut microbiome in a cohort of 38 subjects, including 15 patients with progressive liver disease, 16 patients with non-progressive liver disease, and 7 control subjects. We found that although patients were generally clustered in three groups according to disease status, metabolites showed better separation than microbial species. Furthermore, eight serum metabolites were correlated with two microbial species, among which seven metabolites were decreased in patients with progressive liver disease. Five fecal metabolites were correlated with three microbial species, among which four metabolites were decreased in patients with progressive liver disease. When predicting progressive liver disease from non-progressive liver disease using correlated metabolic and microbial signatures with the random forest model, correlated serum metabolites and microbial species showed great predictive power, with the area under the receiver operating characteristic curve achieving 0.91. The multi-omics signatures identified in this study are helpful for the early identification of patients with progressive alcohol-associated liver disease, which is a key step for therapeutic intervention.
Keywords: gut microbiota; metabolome; multi-omics; random forest.