No abstract

Significant heterosis for seed yield in oilseed rape (Brassica napus L.) has created interest in the development of hybrid cultivars. The objective of this study was to determine the value of restriction fragment length polymorphism (RFLP) markers in predicting hybrid performance. The hybrids and parents of two sets of diallel crosses were evaluated at three environments for seed yield and other agronomic traits. The parents of the first diallel were seven oilseed rape cultivars and the parents of the second diallel were seven unselected $6 lines derived from the cultivars. Genetic distances (GD) between the parents crossed in the diallels were estimated by RFLP data from 43 DNA clones. Both general combining ability (GCA) and GD estimates were significantly (P < 0.05) correlated with hybrid seed yield in both diallels, although GCA was more greatly correlated than GD. Genetic distance was significantly correlated with heterosis for seed yield only in the inbred diallei whereas GCA was significantly correlated with heterosis only in the cultivar diallel. Midparent yield was significantly correlated only with heterosis for the cultivar diallel. A multiple linear regression model that included both the GD and GCA of the parents was more greatly correlated with hybrid seed yield than any variable alone. The GCA values were significantly correlated with hybrid plant height, and seed oil and protein concentration in both diallels whereas GD was significantly correlated only with hybrid plant height. These results suggest that GD estimates alone do not identify high yielding hybrid combinations with the consistency to be useful in breeding programs.

The combined use of doubled haploid lines and molecular markers can provide new genetic information for use in breeding programs. An F1-derived doubled haploid (DH) population of Brassica napus obtained from a cross between an annual canola cultivar ('Stellar') and a biennial rapeseed ('Major') was used to construct a linkage map of 132 restriction fragment length polymorphism loci. The marker loci were arranged into 22 linkage groups and six pairs of linked loci covering 1016 cM. The DH map was compared to a partial map constructed with a common set of markers for an F2 population derived from the same F1 plant, and the overall maps were not significantly different. Comparisons of maps in Brassica species suggest that less recombination occurs in B. napus (n = 19) than expected from the combined map distances of the two hypothesized diploid progenitors, B. oleracea (n = 9) and B. rapa (n=10). A high percentage (32%) of segregating marker loci were duplicated in the DH map, and conserved linkage arrangements of some duplicated loci indicated possible intergenome homoeology in the amphidiploid or intragenome duplications from the diploid progenitors. Deviation from Mendelian segregation ratios (P < 0.05) was observed for 30% of the marker loci in the DH population and for 24% in the F2 population. Deviation towards each parent occurred at equal frequencies in both populations and marker loci that showed deviation clustered in specific linkage groups. The DH lines and molecular marker map generated for this study can be used to map loci for agronomic traits segregating in this population.

No abstract

No abstract