Multiparental QTL analysis: can we do it in polyploids?

Abstract

Many ornamental crops are polyploid or even exist at different ploidy levels. Polyploid QTL analysis tools have been developed in recent years, yet they are limited in the population types they accept. Biparental populations are nowadays being regarded as a limited tool for QTL discovery, as only a limited number of QTLs occurs in an experimental cross and their effects might not be stable across genetic backgrounds. Genome-Wide Association Studies include more genetic diversity but suffer from (hidden) genetic structure and low frequency of QTL alleles. Both factors influence QTL detection and effect estimation, decreasing the sensitivity of QTL analysis. Alternatively, multiparental populations (MPP) can be used, potentially combining multiple QTLs and QTL alleles with known population structure and balanced allele frequencies. Breeding populations of interconnected crosses also constitute a form of MPP and QTLs identified in them might be more applicable to commercial cultivars. To perform QTL analysis in polyploids, mixed models or Bayesian approaches that consider pedigree information are recommended. During the analysis, QTL effects are ideally estimated using Identity by Descent (IBD) alleles (genomic regions that originate from the same ancestor) which can be obtained through haplotype estimation. Although MPPs could thus be a powerful set-up to estimate polyploid haplotypes, a software gap was identified as no current polyploid haplotyping tools are able to utilize MPP pedigree information to obtain haplotypes across an MPP. In order to utilize MPPs to their full extent and expand polyploid QTL analyses to encompass typical breeding populations, new haplotyping tools must be developed.