Gut microbiome features and metabolites in non-alcoholic fatty liver disease among community-dwelling middle-aged and older adults

• DOI: 10.1186/s12916-024-03317-y
• Publication Date: 2024-03-07T15:03:06Z
• Authors: Fangfang Zeng; Xin Su; Xinxiu Liang; Minqi Liao; Haili Zhong; Jinjian Xu; Wanglong Gou; Xiangzhou Zhang; Luqi Shen; Ju-Sheng Zheng; Yu-ming Chen
• Keywords: 0303 health sciences; 03 medical and health sciences; R; 16S rRNA gene sequence; Medicine; Gut metabolites; Machine learning algorithms; Gut microbiota feature; Non-alcoholic fatty liver disease; Research Article

Abstract Background The specific microbiota and associated metabolites linked to non-alcoholic fatty liver disease (NAFLD) are still controversial. Thus, we aimed understand how the core gut impact NAFLD. Methods data for discovery cohort were collected from Guangzhou Nutrition Health Study (GNHS) follow-up conducted between 2014 2018. We 272 metadata points 1546 individuals. input into four interpretable machine learning models identify important with These subsequently applied two validation cohorts [the internal ( n = 377), prospective 749)] assess generalizability. constructed an individual microbiome risk score (MRS) based on identified animal faecal transplantation experiment using samples individuals different levels of MRS determine relationship Additionally, targeted metabolomic sequencing analyse potential metabolites. Results Among used, lightGBM algorithm achieved best performance. A total 12 taxa-related features selected by further used calculate MRS. Increased was positively presence NAFLD, odds ratio (OR) 1.86 (1.72, 2.02) per 1-unit increase in An elevated abundance f__veillonellaceae ) increased NAFLD risk, whereas f__rikenellaceae , f__barnesiellaceae s__adolescentis a decreased Higher microbiota-derived bile acids (taurocholic acid) might be both higher risk. FMT mice confirmed causal association development Conclusions that alteration composition biologically relevant development. Our work demonstrated role