A Systems Approach to Identify Genetic and Environmental Regulators of Metabolism

For more than a century, it has been recognized that our genetic inheritance and our environment interact to shape who we are and how we act (nature vs. nurture€), and the study of genetics has allowed us to explain why traits can vary dramatically between individuals (i.e. trait variance), and yet often be strongly shared within families (i.e. trait heritability). Scientists, statisticians, and physicians can calculate heritability, and can observe how both genes and environment influence health, but relatively few interventions or treatments have stemmed from this expanded knowledge. In large part this is because testing biomedical hypotheses is difficult, because we are all so different, and because environmental factors are so hard to control. In 2001 the first (nearly) complete DNA sequence of the human genome was generated after years of technical development and brute force effort. This accomplishment has ushered in a new field of personalized medicine, as prevention and treatment can both be tailored based on genome and environment. Unfortunately, the science of genomic sequencing has greatly outpaced our ability to actually understand the genetic code, and it remains difficult to make accurate predictions of an individual's characteristics and susceptibilities except in a few clear-cut cases such as eye color or risk of Huntington's disease. There are several reasons for this disconnect: (1) many traits and diseases are driven by complex interactions among environmental causes and genetic risks, (2) there are many aspects of genetics which we do not fully understand and that cannot be easily observed (e.g. non-coding RNAs, epigenetic modifications, and complex metagenomic cohorts that we all carry), and (3) gene products (such as mRNA and protein) change dramatically over time and across cell types in complex and even unpredictable ways. Today, myriad developments promise to improve our capacity to predict resistance and susceptibility to diseases based on our individual genome. This technical capability has also provided new avenues for developing therapeutic agents and/or lifestyle changes. However, no matter our developments in scientific understanding, personalized medicine will never allow perfect prediction—environmental influences and personal choices will always affect our health, and medical treatments will still require information gained from standard medical checkups. Rather than perfect prediction, personalized medicine will instead provide more accurate predictions on health than those previously possible. Thus, those at high risk for (e.g.) diabetes will know better to watch their diet, while those at low risk should remember that biology is never completely predictable. In this thesis, I have analyzed 45 strains of mice from a genetic reference population called the BXD with the goal of identifying major gene regulators of metabolic phenotypes, such as exercise capacity and glucose response. Each member of this BXD family, which contains ~150 distinct but related lines (or strains€), has a unique genetic makeup that has been fixed by inbreeding. Each family member is thus available as an unlimited supply of identical twins€ which may be studied over years and among laboratories. With this population, it is possible to both (1) test what would occur to a single individual in different environmental conditions and (2) analyze how much environmental influences vary across genetically-diverse phenotypes.