Fig. 4

a Posterior probabilities of significant (noted by 95% credible intervals) microbial sub-communities (19, 23, 25, and 26) were fitted against dog allergens using a local regression. b The relative frequency of the significant sub-communities was observed for samples grouped by neighborhood—urban, suburban, and rural. c The posterior probability of frequently co-occurring taxa found within each sub-community was observed at the genus level