The effect of host admixture on wild house mouse gut microbiota is weak when accounting for spatial autocorrelation

Abstract

The question of how interactions between the gut microbiome and vertebrate hosts contribute to host adaptation and speciation is one of the major problems in current evolutionary research. Using bacteriome and mycobiome metabarcoding, we examined how these two components of the gut microbiota vary with the degree of host admixture in secondary contact between two house mouse subspecies (Mus musculus musculus and M. m. domesticus). We used a large data set collected at two replicates of the hybrid zone and model-based statistical analyses to ensure the robustness of our results. Assuming that the microbiota of wild hosts suffers from spatial autocorrelation, we directly compared the results of statistical models that were spatially naive with those that accounted for spatial autocorrelation. We showed that neglecting spatial autocorrelation can strongly affect the results and lead to misleading conclusions. The spatial analyses showed little difference between subspecies, both in microbiome composition and in individual bacterial lineages. Similarly, the degree of admixture had minimal effects on the gut bacteriome and mycobiome and was caused by changes in a few microbial lineages that correspond to the common symbionts of free-living house mice. In contrast to previous studies, these data do not support the hypothesis that the microbiota plays an important role in host reproductive isolation in this particular model system.