RESEARCH PROJECTS

ENVIRONMENTAL MICROBIOMES

We are analyzing part of the already characterized global Tara Oceans metagenome and metatranscriptome dataset and our own partially characterized datasets from estuarine ecosystems that differ in substrate complexity.  These datasets are being analyzed in relation to high resolution organic matter data as well as more traditional measures of ecosystem processes (e.g., bacterial production) collected at the time of sampling.
In addition, microbiomes/metabolomes of bulk soils from the rhizosphere experiment with increasing cover crops described below are also being analyzed for the relationship between functional redundancy and soil health.

HOST-ASSOCIATED MICROBIOMES

Three host-associated ecosystems are being analyzed as part of this project: 1) human gut microbiomes from a sub cohort of men within the Health Professionals follow-up study (BioProject:PRJNA354235) and 2) gut microbiomes of mice fed diets of simple, moderate and high substrate complexity and 3) rhizosphere microbiomes of roots associated with increasing numbers of cover crops that result in increasing substrate complexity. Metagenomics, metatranscriptomics and metabolomics are being used to investigate functional redundancy in relation to substrate complexity in these ecosystems. The mouse and rhizosphere experiments are currently being conducted as part of this project.

DEEP LEARNING TO COMBINE DATASETS

With more and more paired metagenome and metatranscriptome datasets available for microbial communities, it is imperative to develop new methods that capture integrated diversity across datasets. Here, inspired by the recent development of deep generative learning to represent complex data, we are developing new generative representation models for genes, microbial genomes and microbial communities. Then, we are using the new generative representation of the microbial community to predict the measured traits and understand taxonomic and functional diversity and functional redundancy.

FUNCTIONAL REDUNDANCY THEORY

We are building a series of predictive models to explore the mechanistic underpinnings of functional redundancy and the relationship between functional redundancy and ecosystem complexity.  In this context, we are using a metacommunity framework, in which there is a broader, regional taxon pool, and a focal, local community. We are examining how dispersal of taxa from the regional pool results in the assembly of local communities with varying degrees of functional redundancy based on the complexity of the local environment.