Fits statistical models to detect microbial strains associated with a phenotype. For small sample sizes, models can be fitted using empirical Bayes maximum a posteriori (MAP) priors to improve detection power. Supports various experimental design scenarios, including repeated measures.