Population based genotype imputation
Date
Author
Institution
Degree Level
Degree
Department
Supervisor / Co-Supervisor and Their Department(s)
Examining Committee Member(s) and Their Department(s)
Citation for Previous Publication
Link to Related Item
Abstract
In this dissertation, I focus on the study of genotype imputation in population data. Genotype imputation is a process of inferring missing values for genotype data and has been extended to predicting “untyped” genotypes for samples in low-density chips with a reference population assayed using dense marker chips. It has been successfully and routinely applied to merge genotype datasets of different densities that arise from various genotyping and sequencing platforms. First, I examine and compare several influential imputation models that incorporate biological concepts, mine for associations among genetic markers and explore genetic relatedness. I further evaluate the effect of imputation on genomic prediction, which combines dense marker data with phenotypic data for improving quantitative traits. Additionally we propose a multi-step strategy that can work with any existing genotype imputation methods to boost the accuracy of imputation from low-density chips to high-density chips. Finally we describe a new hidden Markov model for genotype imputation based on an existing framework.