Genome-wide variation in human germline and postzygotic mutation rates

Abstract

De novo mutations (DNMs) are new variants that arise in the parental germline or early embryo. In this dissertation, I apply long-read sequencing technology to quads and a multi-generational pedigree to discover DNMs across the genome and quantify the de novo mutation rate. First, I demonstrate that long reads enable DNM discovery in previously inaccessible regions of the genome. These newly accessible regions, largely marked by repetitive sequence, have a significantly higher mutation rate than their unique counterparts, including an approximately 66% enrichment in segmental duplications. I was able to trace the origins of DNMs to either the parental germline or early rounds of embryogenesis, revealing that at least 15% of single nucleotide DNMs arise postzygotically, a 50% increase from earlier studies. Further, I found that 60% of postzygotic mutations are transmitted to the next generation, meaning that they contribute to segregating variation in the population. Finally, I estimate the de novo mutation rate to be approximately 1.2-1.3Ã 10-8 substitutions/base pair/generation for 30 year old parents, and the postzygotic mutation rate to be approximately 0.23Ã 10-8 substitutions/base pair/generation. My analyses reveal that repetitive regions are in fact hypermutable, and that more variation arises postzygotically than previously thought. This work also lays the foundation for the next frontier in DNM discovery: comparing assembled parent and child genomes to reveal variation in the most complex and mutable parts of the genome.