Computational methods to dissect the genetic basis of human disease

Abstract

Genome-wide association studies (GWAS) have been successful in identifying disease-associated genetic variants. However, the path from GWAS to biological insight remains challenging, notably in identifying relevant biological pathways, explaining mechanistic links between diseases, and nominating disease-critical tissues and cell types. In this thesis, I introduce computational methods to dissect the genetic basis of human disease by integrating GWAS with functional data. In the first chapter, I integrate the GWAS with biological pathways and gene networks to elucidate biological mechanisms. I identify significantly associated pathways and highlight the importance of accounting for regulatory annotations in pathway enrichment and gene network analyses. In the second chapter, I investigate the shared genetic architecture between Mendelian disease and common disease by developing a machine learning framework to impute and denoise Mendelian disease-derived pathogenicity scores. I assess the informativeness of Mendelian pathogenicity scores for common disease and improve upon existing scores. In the third chapter, I prioritize disease-critical cell types by integrating GWAS with single-cell gene expression and chromatin accessibility profiling of fetal and adult brains. I show that identified disease-cell type associations recapitulates known biology while informing future analyses of disease mechanisms.