When is microbial strain sharing evidence for transmission?

In humans and other social animals, social partners have more similar microbiomes than expected by chance, suggesting that social contact transfers microorganisms. However, social microbiome transmission can be difficult to identify if social partners also have other traits in common or live in a shared environment. Strain-resolved metagenomics has been proposed as a solution for tracking microbial transmission. Using a fecal microbiota transplant dataset, we show that strain sharing can recapitulate true transmission networks under ideal settings when donor-recipient pairs are unambiguous. However, gut metagenomes from a wild baboon population, where social networks predict compositional similarity, show that strain sharing is also driven by demographic and environmental factors that can override signals of social interactions. We conclude that strain-level analyses provide useful information about microbiome similarity, but other facets of study design, especially longitudinal sampling and careful consideration of host characteristics, are essential for conclusions about the underlying mechanisms.