A genetic linkage map is a valuable tool for quantitative trait locus mapping, map-based gene cloning, comparative mapping, and whole-genome assembly. Alfalfa, one of the most important forage crops in the world, is autotetraploid, allogamous, and highly heterozygous, characteristics that have impeded the construction of a high-density linkage map using traditional genetic marker systems. Using genotyping-by-sequencing (GBS), we constructed low-cost, reasonably high-density linkage maps for both maternal and paternal parental genomes of an autotetraploid alfalfa F1 population. The resulting maps contain 3591 single-nucleotide polymorphism markers on 64 linkage groups across both parents, with an average density of one marker per 1.5 and 1.0 cM for the maternal and paternal haplotype maps, respectively. Chromosome assignments were made based on homology of markers to the M. truncatula genome. Four linkage groups representing the four haplotypes of each alfalfa chromosome were assigned to each of the eight Medicago chromosomes in both the maternal and paternal parents. The alfalfa linkage groups were highly syntenous with M. truncatula, and clearly identified the known translocation between Chromosomes 4 and 8. In addition, a small inversion on Chromosome 1 was identified between M. truncatula and M. sativa. GBS enabled us to develop a saturated linkage map for alfalfa that greatly improved genome coverage relative to previous maps and that will facilitate investigation of genome structure. GBS could be used in breeding populations to accelerate molecular breeding in alfalfa.

Background

Genomic selection based on genotyping-by-sequencing (GBS) data could accelerate alfalfa yield gains, if it displayed moderate ability to predict parent breeding values. Its interest would be enhanced by predicting ability also for germplasm/reference populations other than those for which it was defined. Predicting accuracy may be influenced by statistical models, SNP calling procedures and missing data imputation strategies.

Results

Landrace and variety material from two genetically-contrasting reference populations, i.e., 124 elite genotypes adapted to the Po Valley (sub-continental climate; PV population) and 154 genotypes adapted to Mediterranean-climate environments (Me population), were genotyped by GBS and phenotyped in separate environments for dry matter yield of their dense-planted half-sib progenies. Both populations showed no sub-population genetic structure. Predictive accuracy was higher by joint rather than separate SNP calling for the two data sets, and using random forest imputation of missing data. Highest accuracy was obtained using Support Vector Regression (SVR) for PV, and Ridge Regression BLUP and SVR for Me germplasm. Bayesian methods (Bayes A, Bayes B and Bayesian Lasso) tended to be less accurate. Random Forest Regression was the least accurate model. Accuracy attained about 0.35 for Me in the range of 0.30-0.50 missing data, and 0.32 for PV at 0.50 missing data, using at least 10,000 SNP markers. Cross-population predictions based on a smaller subset of common SNPs implied a relative loss of accuracy of about 25% for Me and 30% for PV. Genome-wide association analyses based on large subsets of M. truncatula-aligned markers revealed many SNPs with modest association with yield, and some genome areas hosting putative QTLs. A comparison of genomic vs. conventional selection for parent breeding value assuming 1-year vs. 5-year selection cycles, respectively, indicated over three-fold greater predicted yield gain per unit time for genomic selection.

Conclusions

Genomic selection for alfalfa yield is promising, based on its moderate prediction accuracy, moderate value of cross-population predictions, and lack of sub-population structure. There is limited scope for searching individual QTLs with overwhelming effect on yield. Some of our results can contribute to better design of genomic selection experiments for alfalfa and other crops with similar mating systems.

A large set of genome-wide markers and a high-throughput genotyping platform can facilitate the genetic dissection of complex traits and accelerate molecular breeding applications. Previously, we identified about 0.9 million SNP markers by sequencing transcriptomes of 27 diverse alfalfa genotypes. From this SNP set, we developed an Illumina Infinium array containing 9,277 SNPs. Using this array, we genotyped 280 diverse alfalfa genotypes and several genotypes from related species. About 81% (7,476) of the SNPs met the criteria for quality control and showed polymorphisms. The alfalfa SNP array also showed a high level of transferability for several closely related Medicago species. Principal component analysis and model-based clustering showed clear population structure corresponding to subspecies and ploidy levels. Within cultivated tetraploid alfalfa, genotypes from dormant and nondormant cultivars were largely assigned to different clusters; genotypes from semidormant cultivars were split between the groups. The extent of linkage disequilibrium (LD) across all genotypes rapidly decayed to 26 Kbp at r(2) = 0.2, but the rate varied across ploidy levels and subspecies. A high level of consistency in LD was found between and within the two subpopulations of cultivated dormant and nondormant alfalfa suggesting that genome-wide association studies (GWAS) and genomic selection (GS) could be conducted using alfalfa genotypes from throughout the fall dormancy spectrum. However, the relatively low LD levels would require a large number of markers to fully saturate the genome.