I will give an overview of methods that use generative probabilistic models to describe microbial community structure as determined through next generation sequencing. I will discuss the concept that the human gut microbiota is derived from a finite number of discrete types or enterotypes. I will show that Dirichlet-multinomial mixtures allow a more nuanced description of enterotypes as diffuse peaks in community configurations. I will then extend the idea of Dirichlet priors for community configurations to hierarchical Dirichlet processes which allow ‘unseen’ species to be modelled. I will use these both as a means for fitting the ecological model, Hubbell’s Unified Neutral Theory of Biodiversity, and for avoiding rarefaction in microbial community diversity estimation.