The environment is perhaps the major contributor to human disease, yet its effect is largely ignored in whole genome approaches to genetics. In this CEGS, we will attack head-on the mechanisms through which environment influences genomic function, focusing on two extremely important exposures highly relevant to human health: diet and its relationship to metabolic disease and cancer; and stress related to neuropsychiatric disease. This requires nearly complete control of the experimental system in a way that cannot be done in humans.
Our first aim i s develop a new foundational experimental mouse model for understanding gene-environment interaction (GxE). We will expose crosses of the genetically heterogeneous Collaborative Cross population of mice to three well-controlled diets - Western, Mediterranean, ketogenic - and measure phenotypes across organ systems representing metabolic and cardiovascular disease. We will also determine the effects of diet on azoxymethane-induced colon tumors; as well as stress on phenotypes relevant to neuropsychiatric disease. We will couple these analyses to whole genome bisulfite sequencing, chromatin accessibility analysis and transcriptomics of target tissues including liver, adipose tissue, colon, hippocampus, prefrontal cortex, and blood, measuring the nature and flow of information among environment, genome, epigenome, and transcriptome in determining phenotype.
Our second aim i s to develop new statistical methods as well as a novel mathematical framework for handling the interaction between genetics, epigenetics, and exposure, which will allow us to model how information is passed between these three domains to ultimately shape phenotype. These methods include causal inference testing, analysis of genomically discontiguous genotype-epigenotype relationships, and novel stochastic approaches based on fundamental concepts of statistical physics and information theory.
Our third aim i s to perform replication in nave animals and in highly relevant human epidemiologic cohorts: DWH and ALSPAC for diet exposures and metabolic disorders; NHSII and EPIC for diet and colon cancer; and ALSPAC and PIRC for stress and behavioral traits.
Our fourth aim i s to develop and promulgate new measurement, analytical, and computational technologies for comprehensive genomic analysis of GxE. These include new biological resources and software packages, as well as biochemical, cellular and computational tools to test and improve the models and conclusions from the other Aims, and of broad general use to the genomics community, including: analysis of single or multiple marks for determining long-range GxE relationships; single cell and in situ analysis of gene expression and epigenetic modifications; and novel computational approaches including ultrafast alignment of large datasets. This highly interdisciplinary proposal involves completely novel combinations of mouse genomics, statistical physics, biostatistics, computer science, biochemistry, epidemiology, and single cell analysis, in order to understand the essential question of how genomic function is shaped by the environment.
The environment is perhaps the major contributor to human disease, yet its effect is virtually impossible to control for in human genetic studies. We are circumventing this problem by modeling in mice two extremely important exposures highly relevant to human health: diet and its relationship to metabolic disease and cancer; and stress related to neuropsychiatric disease. We will integrate a highly relevant genetic mouse model with advanced methods from statistical physics and relate these directly to human disease studies to answer one of the most crucial but vexing problems in genomic medicine, while continuously sharing all of our data and methods with the genomics community for the greatest possible impact for human health.
| Barrington, William T; Wulfridge, Phillip; Wells, Ann E et al. (2018) Improving Metabolic Health Through Precision Dietetics in Mice. Genetics 208:399-417 |
| Konganti, Kranti; Ehrlich, Andre; Rusyn, Ivan et al. (2018) gQTL: A Web Application for QTL Analysis Using the Collaborative Cross Mouse Genetic Reference Population. G3 (Bethesda) 8:2559-2562 |
