Estimating genetic relatedness in polyploid species from sequencing data
Timothy P. Bilton1,2, Matthew R. Schofield1, Ken G. Dodds2 and Michael A. Black3
1. Department of Mathematics and Statistics, University of Otago, Dunedin
2. Invermay Agricultural Centre, AgResearch, Mosgiel
3. Department of Biochemistry, University of Otago, Dunedin
Genetic relatedness refers to the proportion of genetic information between two individuals that is derived from a common ancestor. Computing estimates of genetic relatedness is important in many genetic applications, such as genomic selection, parentage analysis and inferring genetic population structure. Estimators for genetic relatedness have been developed and extensively used in diploid species that have two sets of chromosomes (e.g., humans and most animal species). However, some species (known as polyploids) have a more complex genetic architecture in that they have more than two sets of chromosomes. For these species, relatedness estimation is more complicated, as there are additional genetic phenomena at play, and to date there has been little research in computing relatedness in polyploids. Additionally, estimating relatedness with data generated from the latest sequencing technology is complicated by the presence of errors in the form of missing parental information and incorrectly sequenced bases. These errors, if not taken into account, lead to underestimation of relatedness. We derive theoretical results extending relatedness estimation to polyploids and propose a new estimator that accounts for the errors associated with sequencing data. Using simulated and real data, we examine the performance of these estimators under various scenarios. Lastly, we discuss some limitations of the estimators and scope for future research.
This presentation is eligible for the NZSA Student Prize.