On the robustness of inference of association with the gut microbiota in stool, rectal swab and mucosal tissue samples

The gut microbiota plays an important role in human health and disease. Stool, rectal swab and rectal mucosal tissue samples have been used in individual studies to survey the microbial community but the consequences of using these different sample types are not completely understood. In this study, we report differences in stool, rectal swab and rectal mucosal tissue microbial communities with shotgun metagenome sequencing of 1397 stool, swab and mucosal tissue samples from 240 participants. The taxonomic composition of stool and swab samples was distinct, but less different to each other than mucosal tissue samples. Functional profile differences between stool and swab samples are smaller, but mucosal tissue samples remained distinct from the other two types. When the taxonomic and functional profiles were used for inference in association with host phenotypes of age, sex, body mass index (BMI), antibiotics or non-steroidal anti-inflammatory drugs (NSAIDs) use, hypothesis testing using either stool or rectal swab gave broadly significantly correlated results, but inference performed on mucosal tissue samples gave results that were generally less consistent with either stool or swab. Our study represents an important resource for determination of how inference can change for taxa and pathways depending on the choice of where to sample within the human gut.