Search

As data for microbial community structures found in various environments has increased, studies have examined the relationship between environmental labels given to retrieved microbial samples and their community structures. However, because environments continuously change over time and space, mixed states of some environments and its effects on community formation should be considered, instead of evaluating effects of discrete environmental categories. Here we applied a hierarchical Bayesian model to paired datasets containing more than 30,000 samples of microbial community structures and sample description documents. From the training results, we extracted latent environmental topics that associate co-occurring microbes with co-occurring word sets among samples. Topics are the core elements of environmental mixtures and the visualization of topic-based samples clarifies the connections of various environments. Based on the model training results, we developed a web application, LEA (Latent Environment Allocation), which provides the way to evaluate typicality and heterogeneity of microbial communities in newly obtained samples without confining environmental categories to be compared. Because topics link words and microbes, LEA also enables to search samples semantically related to the query out of 30,000 microbiome samples., Author summary In the past decade, microbiomes from various natural and human symbiotic environments have been thoroughly studied. However, our knowledge is limited as to what types of environments affect the structure of a microbial community. In the first place, how can we define “environments”, in particular, the environmental entities that are often continuously varying and difficult to discretely categorize? We assumed that environments could be represented from microbiome data because the structure of microbial communities reflect the state of the environment. We applied a probabilistic topic model to a dataset containing taxonomic composition data and natural language sample descriptions of >30,000 microbiome samples and extracted “latent environments” of the microbial communities, which are core elements of environmental mixtures. Integrating the training results of the model, we developed a web application to explore the microbiome universe and to place new metagenomic data on this universe like a global positioning system. Our tool shows what kinds of the environment naturally exist and are similar to each other on the perspective of the structural patterns of microbiome, and provides the way to evaluate the commonality and the heterogeneity of users’ microbiome samples.