Association of candidate genes with flowering time and forage quality traits in Lolium perenne

Association mapping is a complementary approach to mapping based on F2 or backcross populations. It uses populations of unknown pedigree to exploit the recombination events that have occurred over many generations, in theory enabling a more refined mapping than is usually possible with conventional mapping families. Many of the challenges with this approach include correction for multiple testing, distinguishing real associations from false ones, and genetic structure in populations. We have used a candidate gene approach to association mapping in perennial ryegrass (Lolium perenne), since this minimises some of these issues. We have used 864 genotypes originating from nine populations to find associations between single nucleotide polymorphisms (SNPs) and flowering time, water soluble carbohydrates (WSC) and other forage quality traits. Some of the alleleic variants in the neutral/alkaline invertase gene LpcAI associated with WSC, nitrogen content (N) and dry matter digestibility (DMD), but none were found in phenotypic data from both years. In contrast, a SNP located 265 bp upstream of the homologue in L. perenne, LpHD1, of the photoperiod control gene CONSTANS in Arabidopsis thaliana and HD1 in rice, was consistently associated with flowering time. We also found a significant association between allelic variants at one locus at the homologue of the HD3a gene in rice (FT or flowering time locus in A. thaliana). We have attempted to verify these associations by crossing selected genotypes from the association mapping population with genotypes of late flowering turf grass varieties. The "late allele" in the LpHD1 locus appeared to delay flowering by 2-3 days. The effect of alleles at the HD3a locus was more difficult to interpret due to the presence of multiple alleles, but "late alleles" did appear to delay flowering by 2-3 days as well. Furthermore, genotyping of populations off or on paths appear to support the effect of the late alleles at these two loci. These results are encouraging for the prospects of association mapping at least for some traits