The feasibility of producing inbred rather than F1 hybrid cultivars in Brussels sprouts: predictions from early generations

Predicting the performance of recombinant inbred lines derived by single seed descent (SSD) following crosses between inbred lines is most reliable for traits in which epistasis and genotype environment interaction are absent. The analysis of two field trials of Brussels sprouts indicated the presence of such effects in seven quantitative characters. These effects could be minimised by transforming the scale on which each character was measured. Predictions were made of the likely performance of the recombinant inbred lines currently being produced by SSD at the Institute of Horticultural Research, Wellesbourne, compared with that of current F₁ hybrid varieties. The predictions from the first trial of F₂ material were encouraging for harvest date, stem length and waste weight, but low for sprout number, quality, marketable yield and total yield. Predictions from F₃ material in the second trial were good for sprout number and total yield (in addition to harvest date, stem length and waste weight as before), but still low for marketable yield and quality, with less than 0 01% of lines expected to perform better than the F₁. Model fitting to variances suggested the presence of linkage disequilibrium in dispersion for waste weight, but no information could be obtained about linkage disequilibrium for the two most important characters, marketable yield and quality. The broad heritability of each character except stem length was lower than 0 5 (0 3 or lower for total yield, waste weight and quality), suggesting that advanced generation selection, as used in SSD, might be more successful than selection in the early generations.