Abstract
In many countries, tomato (Solanum lycopersicum L.) is the most important horticultural species, but good variety seeds are expensive. In hybrid breeding, the number of possible single crosses is too large with only a few lines. With 60 lines, for example, 1770 single-cross hybrids can be formed. This makes it expensive and even impossible to conduct an adequate experimental evaluation. These cases require the availability of methods for predicting hybrid performance. This study was designed to evaluate a method to predict fruit yield in tomato hybrids based on genomic fingerprints, the theory of mixed models, and the experimental evaluation of a sample of hybrids. Ninety seven inbreds were studied with 17 ISSR primers, the Jaccard coefficient and the minimum variance method of Ward. The inbreds were assigned to four groups. Two of them named X and Y were chosen to predict performance, with 5 and 8 inbreds, respectively; their forty intergroup single-crosses were generated, and their fruit yield determined thereafter. The performance of crosses was predicted with the best linear unbiased prediction method (BLUP). With m = 50, 100, 200, 400, 600, 1000 sets of n = 5, 10, 15, 20, 25, 30, it was possible to obtain 40 hybrids. Thirty five of them were randomly selected and the performance of the remaining 40-n crosses was predicted. The correlation coefficients between predicted and measured yields were calculated. The average correlation coefficients fluctuated between 0.448 and 0.792 (α≤0.01). Results suggest that the prediction of hybrid fruit tomato performance might be successful using the method BLUP.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have