Abstract
Actinidia chinensis is a mixed-ploidy species complex of economic significance due to the commercialisation of its fresh fruit which are collectively known as ‘Kiwifruit’. Domestication of A. chinensis varieties began in New Zealand as early as 1904 where selection for superior fruit quality and storage led the development of a substantial export market. Domesticated kiwifruit varieties have undergone numerous selective events related to fruit quality and pest tolerance including the Pseudomonas syringae pv. actinidiae (Psa) outbreak of 2010. The rapid domestication and strong selective pressures applied to kiwifruit germplasm suggests a hypothesis that recent selective events will have had a pronounced affect on allele frequencies associated with loci of effect within the recorded pedigree of breeding programmes. However, the genetic history of kiwifruit varieties is confounded by inbreeding, ploidy manipulation and inter-taxon crosses.
Existing tools for the detection of selective sweeps have primarily been developed for the analysis of wild populations and/or comparative analysis of relatively independent domesticated populations. These methods are known to be biased in the presence of significant co-ancestry or genetic bottlenecks and few of them are readily applicable to polyploid taxa. Within this thesis we aim to develop a range of methods for curating polyploid genotypes, integrating them with pedigree data, and performing genome-wide scans for selection within the recorded history of a mixed-ploidy pedigree.
First we develop Markov chain Monte Carlo based methods for assembling micro-haplotypes from capture-GBS sequencing data at any ploidy level. These methods are designed to generate robust micro-haplotype markers within predefined genomic regions suitable for population genetic analyses. Here we argue that micro-haplotype markers are a preferable marker type to SNPs due to their highly multi-allelic nature which makes them better suited for the analysis of inheritance and selection.
We then explore the application of kinship estimation methods from both pedigree and marker data for validating population structure and sampling within a mixed-ploidy context. Our work demonstrates that the integration of pedigree and marker methods is an effective means of identifying both sampling and pedigree errors. We contrast kinship based sample validation with principal component analysis and apply these tools to validate population structure among advanced selections in a complex A. chinensis pedigree.
In relation to validation of population structure, we apply micro-haplotype markers to the characterisation of unreduced gametes within an A. chinensis pedigree. We use gene-drop simulation to approximate the probability of observed genotypes under competing hypotheses and compare these hypotheses in a Bayesian framework. We demonstrate that this method can be applied even in the absence of sample data for the direct progeny of unreduced gametes. The results of our analysis are indicative of post meiotic restitution of unreduced gametes.
Finally we attempt to identify incomplete selective sweeps within the recorded history of a mixed-ploidy A. chinensis pedigree. We use linear regression of multi-allelic markers to estimate founder contributions to advanced selections and compare these results to the expected contributions approximated by gene-drop simulations. With this approach we identify multiple putative selection loci and discuss the likely genes of effect associated with each locus. In doing so, we demonstrate that it is possible to detect putative signatures of selection within inbred mixed-ploidy pedigrees.