Application of RAD-seq in Evolutionary Genomics of Non-Model Organisms

Abstract

Next generation sequencing (NGS) technologies are revolutionizing how we study genetics and evolution in the modern world. Data is generated at such a fast pace that scientists are struggling to keep up with the innovations in methodology and analytical tools. Genomes are being sequenced at an unprecedented rate, and scientists in fields that until recently found no use in learning molecular techniques are venturing into the world of high-throughput sequencing. Almost 10 years ago, a research group developed Restriction-site Associated DNA Sequencing (RAD-seq), a method that targets polymorphisms in close proximity to restriction cut sites in hundreds of samples simultaneously. The beauty of RAD-seq lies in that it is highly customizable and it does not require a reference genome, or intimate prior knowledge of the genetics of the study organism one would like to use. The most exciting part about new RAD-seq methods being developed is that their accessibility has opened the door to many non-model organisms to be used in new areas of research. The overarching theme of my dissertation is the application of RAD-seq data to answer questions in evolutionary, quantitative, or population genetics and genomics using non-model species. A secondary goal is the development of genomic resources for non-model organisms. In Chapter 1, I studied the genetics of a recent shift from self-incompatibility to self-compatibility in an insular lineage of Tolpis, with an aim to identify putative genomic regions responsible for this shift in mating system. To do that, I assembled a draft genome, annotated it, and used RAD-seq data from a mapping population to discover variants. In my second chapter, I focus on a pyralid moth and the genetic basis of male song characters that are attractive to females. For that purpose, I again used RAD-seq data from hundreds of individuals, and, additionally, I assembled and annotated a genome for this non-model organism. In my last chapter, I focused on the bioinformatic challenges associated with RAD-seq data. I explored the question of whether or not using a genome sequence helps in the construction of loci from RAD-seq reads. The evaluation of the last question is on a fairly basic level but it opens up future avenues that I am excited to explore.