Bio: Microbiome Community Dynamics and Complexity

Marker gene and whole genome shotgun (WGS) sequencing approaches have been utilized to characterize the compositional and functional diversity of microbial communities, which we call the microbiome or the metagenome respectively. Such datasets typically result in multivariate abundance measurements over hundreds to thousands of bacterial or gene types. An experiment in metagenomics may typically encompass measurements over a number of time points. Typically, additional measurements of the state of the host organism, such as host developmental state (weight, height, fat mass), gene expression, inflammation status, metabolite levels, etc., are available on the same or different measurement schedule during the experiment. The analysis goal of such studies is to relate the additional measurements with microbiome dynamics and case-control or treatment status.

This group will use data from the NIH human microbiome project (HMP) and other public datasets for applied statistical analysis and as inspiration for the development of new analysis methods and guidelines for the use of existing methods. Human microbiome studies aim to understand the role of the bacterial communities in physiology and disease.

Specific datasets may include:

(i) Integration of 16S rRNA metagenomic, lipidomic, and immunoproteomic data. Other multi-omic analyses based on the public release of NIH HMP 2 data may be possible, some of which may be longitudinal measurements.

(ii) Data from HMP 1 includes samples from ~240 subjects at 15-18 body habitats each. Metagenomic data are publicly available; accompanying clinical data are available on dbGaP and requires an extensive approval process. Data from repeat visits are available.

(iii) Other publicly available microbiome data and clinical data on dbGaP may also serve as the basis for other investigations; e.g., best practices for analysis of longitudinal multi-omic data, visualization of longitudinal multi-omic data, or standards for metabolomics data.

(iv) Data from mock communities for quantifying and counteracting bias in 16S rRNA surveys. The use of mixture experiments and backward-fitting models to predict “true” community composition. Analysis of existing mixture experiment data, design of future experiments, and development of models for prediction.