Abstract
Candidate varieties must satisfy the requirements of Distinctness, Uniformity and Stability (DUS) based on morphophysiological traits. Breeders of major perennial forage crops strive to achieve statistically significant distinctness of candidate varieties, especially with respect to registered varieties bred from the same genetic base. This study aimed to compare morphophysiological versus molecular distinctness for 10 potential alfalfa varieties selected from the same genetic base and one recently registered variety selected from a similar genetic base. We also aimed to support the optimization of alfalfa molecular distinctness by comparing (a) genotyping-by-sequencing (GBS) versus alfalfa DArTag panel markers, and (b) four statistical criteria. The 11 populations (one originating from mass selection, nine from progeny-based selection, and one from clonal selection) were evaluated for 10 morphophysiological traits ordinarily used for DUS in Italy and were characterized molecularly using three bulked DNA samples of 200 independent plants each per population. Only four morphophysiological traits exhibited significant (P < 0.01) population variation, mainly because of population similarity for autumn dormancy, growth habit, and plant vigor. Morphophysiological distinctness emerged for 20 of the 55 paired comparisons between the 11 populations, with no population showing distinctness from any other. Various GBS and DArTag marker configurations achieved a complete distinctness of each population from any other (at P < 0.01) by statistical criteria based on a principal components analysis of allele frequencies followed by analysis of variance or discriminant analysis of population principal component scores. These criteria showed greater population discrimination than cluster analysis with bootstrap values. Analysis of molecular variance was ineffective for population distinctness, probably because of insufficient number of bulked DNA samples per population. Mantel's test indicated high correlation for Euclidean distances of the populations between GBS and DArTag markers (r = 0.94) and no correlation of these distances with those based on morphophysiological traits. Variety registration on the ground of distinctness from the already registered variety could be granted to only four of 10 selections according to morphophysiological distinctness, and to all selections according to molecular distinctness. Our results support molecular distinctness as a sensitive, quick, and inexpensive tool for alfalfa variety registration, forensic analyses, and control of certified seed.