A platform for mining, visualization and design of microbial interaction networks

Boston University (Charles River Campus), Boston MA

Investigators

Linked publications & trials

Abstract

Project Summary One of the burning questions in the study of the human microbiome is whether and how it is possible to design specific strategies for rebalancing the taxonomic and functional properties of human-associated microbial communities, triggering the transition from ?disease states? to ?healthy states?. While empirical studies provide strong support for the idea that we may be able to cure, or at least treat, a number of diseases by simply transplanting microbiomes, or inducing changes through taxonomic or environmental perturbations, to date little mechanistic understanding exists on how microbial communities work, and on how to extend microbiome research from an empirical science to a systematic, quantitative field of biomedicine. We propose here to establish a computational platform-- a database (Aim 1) with fully integrated analytical software (Aims 2 and 3) --- developed for and with the cooperation of the scientific community. The resource goes beyond cataloguing microbial abundances under different condition; its aim is to enable an understanding of networks of interacting species and their condition-dependence, with the goal of eventually facilitating disease diagnosis and prognosis, and designing therapeutic strategies for microbiome intervention. Our project is centered around three key aims: 1. The creation of a Microbial Interaction Network Database (MIND), a public resource that will collect data on inter-species interactions from metagenomic sequencing projects, computer simulations and direct experiments. This database will be accessed through a web-based platform complemented with tools for microbial interaction network analysis and visualization, akin to highly fruitful tools previously developed for the study of genetic networks; the database will also serve as the public repository of microbial networks associated with human diseases; 2. The implementation of an integrated tool for simulation of interspecies interactions under different environments, based on genomic data and whole-cell models of metabolism; 3. The implementation of new algorithms for microbial community analysis and engineering. These algorithms, including stoichiometric, machine-learning and statistical approaches will facilitate a ?synthetic ecology? approach to help design strategies (e.g. microbial transplants or probiotic mixtures) for preventing and targeting microbiome-associated diseases. Our work will fill a major gap in current microbiome research, creating the first platform for global microbial interaction data integration, mining and computation.