USING DNA MARKERS TO SPEED UP PROGRESS IN SUGARCANE BREEDING

FOR THE PAST 50 or so years, rates of improvement in breeding value of parents in sugarcane breeding programs worldwide appear to be slow, which undoubtedly has impacted adversely on performance of the best clones released to industry. Likely reasons include the high proportion of genetic variation present as complex, non-additive genetic effects in sugarcane for key traits and long generation intervals (typically 8–15 years). The high non-additive genetic variation in modern breeding populations means that phenotypes of a set of clones only poorly predict relative breeding values. It has been hypothesised that DNA markers may be useful in addressing these problems. In this paper two example sets of experiments and results are presented, namely an association mapping study and a study to pilot the use of DNA markers for parent selection. It was shown that markers associated with cane yield and CCS can be readily found and that interactions between markers appear important. It was also shown that markers can predict breeding value better than phenotype. The results and other considerations support the hypothesis that DNA markers could be used in speeding up parental improvement in sugarcane breeding programs, and a scheme to achieve this is proposed. A marker-assisted breeding program could use an approximately 3-year generation interval rather than an 8–15 year cycle, while also predicting breeding value of clones more accurately than current methods. However, application of DNA markers to predict genetic value (for selection of clones for direct commercial use, as opposed to selection of clones as parents) will probably depend on models that can incorporate non-additive (interaction) marker effects.