Plant breeding is undergoing a major change in methodology with the application of DNA-based markers for selection. Little has been done to quantify the efficiency of apple breeding, or indeed tree fruit breeding, with or without the new technology. This study aims to investigate the impact on different optimization criteria of alternative designs for an evaluation of selection candidates established in phase 2 field trials of the Washington State University Apple Breeding (WABP) program using two metrics. Acceptance probability, an extension of the least significance difference (LSD) common in statistics, has been used to evaluate alternative trial designs. Acceptance probability expresses the observed percentage difference a candidate must have when compared to a standard(s) to reject the hypothesis that the mean of the candidate is equal to the mean of the standard(s). Response to selection was also used to compare alternative designs which differ in trial size. This is because variation in numbers of candidates may affect both variation of predicted cultivar mean (similar to the account in LSD and acceptance probability) and the intensity of selection. Apple breeding is a costly and long-term investment (18 years minimum from cross to release), for which improvements in efficiency have huge positive repercussions. Opportunities for improvement identified in the WABP can then be adopted by other tree fruit breeding programs, resulting in faster development of new varieties.
