Prediction of heterotic groups and hybrid performance in South African sunflower (Helianthus annuus L.) germplasm using SSR analysis

English: Breeders would prefer to be able to predict the outcome of crosses prior to the production and testing of lines derived from them in field trials. One way to do this would be to find a correlation between the genetic distances of inbreds and relative yield obtained from the hybrids. The aim of this study was to determine whether the SSR based dendrogram can be of value as a predictor for the best performing combinations between A- and R-lines in the context of South African germplasm as well to determine whether correlations existed between oil content and relative yield and genetic distance and oil content. The study analysed 93 inbred lines, consisting of 49 R-, 40 A- and four B-lines which were planted in a glasshouse at Greytown, KwaZulu-Natal (South Africa). Two experiments were performed on the sets of lines, namely SSR analysis to establish genetic distance and a dendrogram and the second was to establish yield trials with a subset of the lines. A dendrogram was constructed using 55 SSR markers widely distributed over the entire sunflower genome. The objective was to establish genetic distances of the lines involved in order to determine heterotic groups in the hybrid breeding programme. SSR cluster analysis separated lines into two main groups, one included A- and B-lines (females) and the other R-lines (males). The groupings represented the breeding history and basic heterotic pattern of sunflower. Genetic similarities were lower overall for female (A) line x restorer (R) line crosses than for A x A or R x R. The highest level of dissimilarity was found between A9 and A40 and the second highest between A23 and R8. The lowest level of dissimilarity was found between R14 and R41 and the second lowest between R42(HO) and R43(HO). The second experiment included crossing of 33 A-lines to 11 R-lines to produce FI hybrid seed. However, not all crosses were successful. The successful hybrids were planted as lattice designs in a total of 133 trials on six locations. The yield of the hybrids was calculated and expressed relative to the yield of the commercial hybrid PAN 7351 which was included in all trials. Relative yield and oil percentage was calculated. The line x tester analysis was used to determine the combining ability of the inbred lines and to determine if there existed correlations between genetic distance and relative yield, between genetic distance and oil percentage and between relative yield and. oil percentage. The highest yielding combination was A25/R34. Low to no correlations were found between genetic distance and relative yield, between genetic distance and oil percentage and between relative yield and oil percentage. Exceptions were significant correlations found between relative yield and oil percentage for the A-lines crossed to R32, for genetic distance versus relative yield and for relative yield versus oil percentage for the A-lines crossed to R48 and for relative yield versus oil content for the A-lines crossed to R44(RM). The female line A16 exhibited high oil content in most of the combinations it was involved in. The average genetic distance for the lines involved in the yield trials was 0.24. This indicated low differences between the lines used and a possible reason for the low correlations found between combinations. Differences were found in GCA effects with A34 being the best general combiner for relative yield. The combinations A6/R34 and A25/R34 were the combinations with the best SCA effects for relative yield. According to the GCA:SCA ratio the SCA was greater which indicated non-additive gene action. Narrow-sense heritability was found to be 52.98% for relative yield. This was caused by equal parts additive effects and dominant gene actions.