Project 451822

Many common diseases, such as diabetes and heart disease, are thought to be caused in part by the differences in the DNA sequences of our genomes. Each individual has a unique genetic makeup, and many locations in the genome vary between people. The unique genetic variation within each person affects their disease risk in ways we do not completely understand. However, we do know that disease-associated genetic variation often lie in the regions surrounding genes. These are thought to contribute to disease by changing when and where the adjacent genes are activated, for instance, by turning the genes on or off at the wrong time or in the wrong cell. Since we do not understand the code that cells use to interpret the DNA, we lack the ability to predict how genetic variation alters when a gene is activated. Our work aims to measure how cells interpret millions of synthetic pieces of DNA that we introduce in a lab, and then use computer models to learn the code the cells are using to interpret these sequences. Here, the synthetic DNA we will test will be random sequences because we can produce and test it in extremely high throughput, producing data at the scale needed to learn complex computer models that capture the gene regulatory code. We are particularly interested in heart disease and will study the cells that contract the heart as it beats (cardiomyocytes). Once we have learned this code, we will use it to determine how genetic variation alters disease risk. For instance, we think that our work will shed important light on how genetic variants at different locations in the genome interact in ways that disproportionately affect disease risk. This work will improve our ability to identify people at a risk of developing disease, and will enable the development of therapeutics that treat disease.