In addition to the processes structuring free-living communities, host associated microbial communities, so-called microbiomes, are directly or indirectly shaped by the host. This results in a hierarchical data structure where samples are nested under one or several variables representing host-specific features, often spanning across multiple levels of biological organization. Previous statistical tools did not accommodate for this nested data structure, therefore cannot explicitly account for the effect of host-specific features on structuring the microbiome. Acknowledging the multivariate nature of species assemblages, joint species distribution modelling is an extension of generalized linear mixed models, simultaneously drawing on the information from multiple species, revealing community-level responses to how species responds to their environment. I will introduce two models which are under development. Key components of these models are the inclusion of latent factors serving multiple purposes, such as accounting for unmeasured covariates, dimensionality reduction, and model visualization.