Direct measurement of gene-environment interactions by high-throughput precision genome editing

点击翻译按钮获取中文摘要

Abstract Modern genetics has identified many genetic variants that affect traits such as height, but most phenotypic variation still cannot be explained by these variants alone. Importantly, differences in environment often result in individual variation of traits—including disease risk and drug response—for different genotypes. These relationships are known as genotype by environment (GxE) interactions. For example, the sickle cell allele of hemoglobin S causes sickle cell anemia, but also provides a fitness advantage in the presence of malaria by conferring resistance to infection. However, there are few examples where the exact causal variants are known. Therefore, we need to develop new methodology for identifying more of these GxE interactions, to improve prediction of disease risk and treatment outcomes. In this study, I will fill in the gap of knowledge in GxE interactions by establishing an experimental framework for identifying hundreds of causal GxE variants in parallel, providing the first comprehensive view of GxE causal variant landscape. Specifically, I will study how thousands of genetic variants between a laboratory yeast strain (BY) and a vineyard strain (RM) lead to their differences in growth upon stress and drug treatments, as one form of GxE interaction. In Aim 1, I will use a novel gene-editing technology that can detect the fitness effects of thousands of variants in one experiment, as shown in a pilot experiment. Using this method, I will be able to map hundreds of casual variants that contribute to growth differences under various conditions, such as carbon source, oxidative stress and drug treatment. In Aim 2, I will measure allele-specific mRNA expression (ASE) from BYxRM F1 hybrids in above-mentioned conditions and associate the mapped causal GxE variants, to identify GxE variants that influence growth rate through gene expression. Then, I will apply a machine learning model to predict causal GxE genes using the molecular features found in this study. By mapping causal GxE variants, linking them to gene expression and predicting causal genes through gene expression, I will establish a complete framework for accelerating the discovery of GxE interactions.