Elhanan Borenstein, Ph.D.

Dept of Genome Sciences
University of Washington

3720 15th Ave NE
Foege Building, S103B
Box 355065
Seattle, WA 98195-5065

Phone: (206) 685-8165
Fax: (206) 685-7301

The vast majority of microbial species on earth live as part of complex, highly diversified and largely uncharted communities. Of special interest is the human microbiome - the set of microbial species that live inside and on the human body - and its effect on human health. Novel metagenomic studies have revealed marked variations in the composition of the human microbiome in various disease states and during development. Yet, a system-level understanding of the human microbiome, its assembly, metabolic capacity, and effect on the host is still lacking.

To address this challenge, we combine computational systems biology, modeling, integrative analysis, machine learning, and complex networks analysis with meta-omic data, to study the human microbiome as a complex ecosystem, going beyond comparative analysis of the set of species comprising the microbiome or the set of genes they encode. This novel system-level approach is crucial to resolving fundamental questions concerning the human microbiome and its role in human health, with numerous medical applications.

Research in the lab is multidisciplinary in nature and spans several levels of abstraction, ranging from state of the art computational analysis of complex networks and high-throughput data to theoretical studies of mathematical and computational models.

Specific research themes include:

Metagenomic systems biology and metabolic modeling of the human microbiome

Recent years have witnessed exciting advances in metagenomic studies, characterizing a plethora of microbial communities across a wide range of environments (including the human body). These complex communities are shaped by ecological and evolutionary processes and in turn have a huge impact on their environments or hosts. We are developing predictive in-silico metabolic models that will allow us to systematically characterize the relationship between the composition of a microbiome, its environment, and its function. Such models can be applied to predict the complex web of metabolic dependencies between species comprising various microbial communities and the community-level interface between microbial 'super-organisms' and their surroundings. These models will provide a framework for studying various microbiomes of interest as well as designing de-novo communities. Using such models we will further examine the resilience of various communities and the predicted effects of specific perturbations, facilitating, for example, the design of therapeutic targeted manipulations of the human microbiome.

Methods for cross-meta-omic analysis of microbiome-derived data.

Systems thinking, analysis, and modeling are not limited to the study of interactions among various components in the microbiome but should also be applied to studying the links between different facets and measures of the microbiome. Much of the effort in modern microbiome research focuses on generating multiple types of omic data, including, most notably, metagenomics, metatranscriptomics, metaproteomics, and metametabolomics. We aim to develop computational methods for integrating these meta-omic datasets, going beyond the identification of statistical associations and putting forward a systems-level framework that links such data through a comprehensive mechanistic understanding of the microbiome.

Computational metagenomics and analysis of taxonomic and functional variation across health and disease.

In collaboration with multiple groups, we are studying taxonomic and functional shifts in the human microbiome that are associated with a plethora of disease states. We specifically utilize our expertise in shotgun metagenomics processing and annotation, machine learning, statistical analysis, and modeling to provide a rigorous and accurate profiling of the microbiome and to reliably identify specific taxa, pathways, and functional modules that may be linked to disease.

Large scale computational study of complex biological and ecological networks

Biology is all about interactions and many biological and ecological systems are best represented as networks of interacting components, whose structures and topologies are an important determinant of system function and dynamics. We study the topological properties of complex biological networks, with an emphasis on metabolic networks and the evolutionary forces that forge their structure. A modular organization, robustness of the phenotype in the face of genetic and environmental perturbations, and an endless capacity for innovation seem to be major organizing principles of biological systems. We are specifically interested in the interplay between the organization of various networks and the environments in which they evolved and prevail. This 'reverse-ecology' approach provides a promising toolbox to fully characterize the dynamics governing the evolution of networks and the adaptation of organisms and ecosystems to their environmental niches.