Back

Giant bacteria in gut microbiomes

Lyu, Y.; Gruener, M.; Klein, A.; Zoccaratto, L.; Daya, A.; Chang, H.-D.; Bizic, M.; Ionescu, D.

2026-10-01 microbiology
10.64898/2026.09.30.750856 bioRxiv
Show abstract

The advent of sequencing-based omics has transformed our understanding of the gut microbiome, enabling microbial communities to be characterised at unprecedented scale and resolution. However, these approaches provide limited information about cellular traits such as size and morphology. Giant bacteria remain largely overlooked in gut microbiome research, despite their extraordinary cell size, extreme polyploidy and unusual reproductive strategies. Epulopiscium, one of the largest known heterotrophic bacteria, has been studied primarily as an intestinal symbiont of surgeonfish, leaving its broader host range unresolved. Here we show, by a large-scale analysis of publicly available 16S ribosomal RNA amplicon samples, alongside locally collected human samples, that Epulopiscium spp. occur across diverse hosts, including humans and a broad diversity of animals. Full-length 16S rRNA gene phylogenies resolved multiple Epulopiscium-related lineages with distinct host compositions. Among human samples, Epulopiscium sequences were most abundant during the first year of life. In adults, abundance varied across disease states, increasing in inflammatory bowel disease and colorectal cancer while decreasing in neurological conditions compared with a healthy adult baseline. Similar patterns were also observed in mouse models. Our study broadens the known host range of Epulopiscium and establishes giant bacteria as an overlooked component of gut microbiome diversity. These findings highlight the need to combine sequence-based studies with classical microbiological techniques to identify organisms with unique traits, otherwise hidden by taxonomy or phylogeny.

Matching journals

The top 6 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.