Kipoi Seminar - Irene Kaplow (Duke University)

Relating enhancer genetic variation across species to complex phenotypes using machine learning Abstract: Advances in genome sequencing and alignment have enabled us to map enhancers—non-protein-coding regions that control when and where genes are expressed—across hundreds of species. However, widely used nucleotide-level conservation measures cannot be applied to many enhancers, where the conservation levels of individual nucleotides are often low even when enhancer function is conserved and where activity is tissue-specific. To overcome these limitations, we developed the Tissue-Aware Conservation Inference Toolkit (TACIT), in which machine learning models learn the regulatory code connecting genome sequence to enhancer activity in a tissue of interest, allowing us to accurately predict cases where differences in genotype across species are associated with differences in enhancer activity in that tissue. We established a new set of evaluation criteria for machine learning models developed for this task and used these criteria to compare our models to models trained using different negative sets and to conservation scores. We then developed a framework for connecting these predictions to phenotypes in a way that accounts for the phylogenetic tree. When applying our framework to motor cortex to study the evolution of brain size, we identified dozens of new associated enhancers, including enhancers near genes implicated in microcephaly and macrocephaly.