Bacterial metabolic signatures in MASLD predicted through gene-centric studies in stool metagenomes

Metabolic dysfunction-associated steatotic liver disease (MASLD) is a multifactorial condition in which the gut microbiome (GM) plays a central role. However, taxonomic associations derived from 16S ribosomal RNA (rRNA) gene studies have yielded inconsistent results, likely due to limited resolution and functional redundancy across taxa. We aimed to identify robust, functionally relevant microbial markers of MASLD using metagenomics and gene-centric profiling. We analyzed 554 fecal metagenomes from three independent cohorts. Sequencing reads were quality-controlled and taxonomically profiled with multi-marker gene resolution. We quantified the abundance of over 50 target gene families involved in butyrate, methane, trimethylamine (TMA) and short-chain alcohol (SCAs, i.e., ethanol and propanol) metabolism. Their presence was also determined across complete GM genomes and plasmids. Genes involved in butyrate and methane production tended to show lower abundance in MASLD, particularly in cirrhosis, while TMA- and SCA-producing genes were frequently enriched. These functional shifts were accompanied by the depletion of Agathobacter rectalis. Many of the altered genes were highly accessory and encoded on plasmids, suggesting genome-specific functional divergence driven by horizontal gene transfer. MASLD is characterized by a shift toward alcohol- and TMA-producing metabolism, alongside reduced butyrate and methane production -changes driven by accessory and plasmid-borne genes. Gene-centric and mobile genetic element-aware profiling reveals mechanistic microbial contributions to MASLD that remain undetected by taxonomy-based approaches, offering new targets for diagnosis and intervention. • MASLD is linked to consistent, gene-level functional shifts in the gut microbiome. • Genes involved in butyrate and methane synthesis are depleted in MASLD patients. • Genes encoding microbial ethanol, propanol and TMA production are enriched. • Many of these metabolic genes are plasmid-encoded, revealing strain-level divergence. • Functional changes occur independently of major taxonomic shifts.