Early prediction of liver disease using conventional risk factors and gut microbiome-augmented gradient boosting

Gut microbiome sequencing has shown promise as a predictive biomarker for a wide range of diseases, including classification of liver disease and severity grading. However, the potential of gut microbiota for prospective risk prediction of liver disease has not been assessed. Here, we utilise shallow gut metagenomic sequencing data of a large population-based cohort (N=>7,115) and [~]15 years of electronic health register follow-up together with machine-learning to investigate the predictive capacity of gut microbial predictors, individually and in conjunction with conventional risk factors, for incident liver disease and alcoholic liver disease. Separately, conventional and microbiome risk factors showed comparable predictive capacity for incident liver disease. However, microbiome augmentation of conventional risk factor models using gradient boosted classifiers significantly improved performance, with average AUROCs of 0.834 for incident liver disease and 0.956 for alcoholic liver disease (AUPRCs of 0.185 and 0.304, respectively). Disease-free survival analysis showed significantly improved stratification using microbiome-augmented risk models as compared to conventional risk factors alone. Investigation of predictive microbial signatures revealed a wide range of bacterial taxa, including those previously associated with hepatic function and disease. This study supports the potential clinical validity of gut metagenomic sequencing to complement conventional risk factors for risk prediction of liver diseases.