epsSMASH uncovers exopolysaccharide biosynthetic gene clusters in environmental and human microbiomes

Biofilms represent the default mode of bacterial life in natural and built environments, with extracellular polysaccharides (exoPS) serving as essential structural and functional components of the biofilm matrix. Despite their importance, exoPS production in these environments is largely unknown. Here we present epsSMASH, a bioinformatic tool and web service for predicting known and novel exoPS biosynthetic gene clusters (BGCs) in bacterial genomes. Benchmarking showed that comprehensive detection of exoPS gene clusters requires highly contiguous high-quality genome assemblies. We applied epsSMASH to high-quality bacterial genome catalogues representing four major ecosystems: Human gut, soil, ocean and activated sludge from wastewater treatment systems. In all catalogues, epsSMASH identified exoPS BGCs in most genomes (52.8-85.4%), with a median of 1-2 exoPS BGCs per genome. The number of exoPS BGC per genome was highly variable, with some taxa containing up to 19 distinct exoPS BGCs. Pel BGCs were abundant in human gut, ocean and activated sludge microbiomes, and were detected in 14 different phyla, making it the most phylogenetically widespread BGC in these environments. The vast majority (62-96%) of detected exoPS BGCs were uncharacterised. By constructing gene cluster families from uncharacterised systems, we identified novel and phylogenetically widespread exoPS BGCs. We investigated a novel exoPS gene cluster from the activated sludge microbiome and showed that it is conserved in most genera within the order Sphingomonadales. Our results highlight the remarkable number of uncharacterised exoPS gene clusters in environmental microbiomes and establish epsSMASH as an effective tool for identifying and classifying novel exoPS systems.