Meiotic recombination is a fundamental biological mechanism that transfers genetic material between homologous chromosomes. The knowledge of how recombination levels vary across a genome is crucial for the design of efficient association mapping studies as well as evolutionary inference studies, and is also of interest from the point of view of basic molecular biology. Since direct measurement of recombination fraction from experiments is technically challenging and genetic maps cannot provide fine-scale resolution, indirect estimation by applying population genetic methods to Single Nucleotide Polymorphism (SNP) data has proven to be a useful alternative.;Current meiotic models allow for two different kinds of recombination events called "crossing-over" and "gene-conversion". Crossing-over refers to the reciprocal exchange of large chromosomal fragments while gene-conversion refers to the transfer of short DNA tracts that are not accompanied by crossing-over. In this dissertation, we develop a novel method for jointly estimating both crossing-over and gene-conversion rates from population genetic data using summary statistics. Summary based methods are attractive in practice because these are extremely fast, flexible, and can directly provide confidence intervals for the estimated values.;The performance the new method was first tested on simulated datasets and compared with that of the pairwise composite likelihood method (Hudson 2001), that is widely used currently. Then, we applied our method to large human datasets resequenced by Perlegen Sciences and estimated recombination rates along the genome. In addition, we independently analyzed an Arabidopsis thaliana dataset sequenced by the Nordborg lab.;Our results with human data suggest that gene-conversion occurs frequently in the human genome relative to crossing-over. We obtained qualitatively similar results for Arabidopsis thaliana. Both human and Arabidopsis datasets seemed to deviate significantly from models where the ratio of conversion to crossing-over was uniform along the sequence. Over Mb scales, population genetic estimates of recombination rate in humans were strongly correlated with the estimates obtained from genetic maps.
展开▼
