Mapping genetic variation in enzyme velocity to growth rate phenotype

点击翻译按钮获取中文摘要

PROJECT SUMMARY/ABSTRACT Over the last thirty years, our capacity to collect genome sequence information has rapidly outpaced our ability to analyze and interpret it. Despite significant efforts to quantitatively relate genotype to phenotype, we struggle to predict classic Mendelian traits like height from human genetic data. In apparently simpler organisms, such as bacteria, we are often unable to predict the effects of even single mutations on growth rate. Indeed, extending our current knowledge of genotype to the understanding, prediction, and control of global cellular behaviors (phenotype) remains a central goal of biology. This problem is made complex by three factors: 1) the mapping between a single gene’s activity and phenotype is non-linear and generally unknown, 2) the mapping is shaped by epistatic interactions between genes, and 3) the mapping is influenced by environmental factors. Given knowledge of the parameters governing these three relationships, we then need a strategy to combine these data into a quantitative model of phenotype. The goal of this grant is to develop exactly such a strategy, by focusing on an experimentally powerful and well defined instantiation of the genotype to phenotype problem: how variation in metabolic enzyme activity influences the growth rate of a unicellular organism (E. coli). We propose a modeling approach in which epistatic relationships between genes and the environment can all be measured and modeled as continuous, dose-dependent phenomena. To parameterize and test this model, we will collect over 100,000 growth rate measurements sampling genetic and environmental variation in folate metabolism, a well-conserved pathway with important roles in human health and disease. These data will be generated using new methodology developed by my laboratory that combines CRISPR interference (CRISPRi), next generation sequencing, and continuous culture to quantitatively measure growth rates for thousands of mutants in parallel under prescribed environmental variation. A small subset of the growth rate data will be used to mathematically constrain our model (~10-20%), and we will evaluate model performance on the remainder. We will also assess the capacity of the model to predict growth rates for higher order combinations of enzyme activity and environmental perturbations not included in the original data set. At completion, we will have established and tested a complete genotype-phenotype mapping relating changes in folate pathway enzyme activities to growth rate. This final model will be of immediate relevance for understanding how variation in folate metabolic enzymes interacts with environmental conditions to influence resistance to common antibiotics (e.g. trimethoprim). More generally, the modeling framework can be translated to map genotype-to-phenotype relationships in other biochemical pathways and cell types. This approach will provide a new strategy for the engineering of biosynthetic pathways, designing personalized therapies, and inferring the growth rate effects of mutations in higher organisms.