Improving seed yield in red clover through marker assisted parentage analysis

Abstract

Red clover (Trifolium pratense L.) is an important perennial forage crop in Europe. Unfortunately, seed production of modern cultivars is often unsatisfactory and breeding progress for higher seed yield is slow. We evaluated whether marker assisted parentage analysis could identify progeny plants with two high seed-yielding parents in a diploid selection trial. Furthermore, we studied if breeding for high seed yield could be speeded up by selecting progeny plants with two high seed-yielding parents, compared to traditional selection on the mother plant only. In this study, 111 genotypes with excellent vegetative traits from a first cycle selection trial were allowed to pollinate each other in isolation. After seed harvest, ten high seed-yielding plants were identified and their seeds were used to grow ten half-sib progeny populations, with 1,121 plants in total. Two multiplex primer sets targeting 18 SSR loci were designed and used to fingerprint parental and progeny plants. Parentage was reliably determined for 1,083 progeny plants: 135 plants were the result of a cross between two high seed-yielding parents. In a 2-year seed yield trial, the seed yield of these 135 progeny plants was compared to that of a control group of 216 progeny plants from the same ten high seed-yielding mother plants but with no selection on the male contribution, as is current practice. Compared to the control group, progeny plants with two high seed-yielding parents had significantly higher seed yields, 23.0 and 75.9 % higher in 2012 and 2013 respectively, and a significantly higher thousand seed weight, 6.6 and 5.7 % higher in 2012 and 2013 respectively. Seed yield was strongly correlated with flower abundance (r = 0.643) and plant volume (r = 0.593 in 2012 and r = 0.509 in 2013) and negatively correlated with susceptibility to mildew (r = −0.145). We conclude that breeding programs aiming at increasing seed yield in diploid red clover cultivars would strongly benefit from the integration of marker assisted parentage analysis to identify those progeny plants that are the result of a cross between two high seed-yielding plants. This goal cannot be achieved with traditional family selection where only the seed yield of the mother plant is known.