RELATIONSHIPS BETWEEN CLONAL FIELD PERFORMANCE ESTIMATES AND THEIR MID-PATENT BREEDING VALUES FOR THREE PRODUCTIVITY TRAITS

ESTIMATED breeding values (EBV) derived by BLUP methodology on family trial data sets have been used to select parents in BSES breeding programs. It was expected that parents with higher EBVs in crop traits should produce progeny with good field performance. EBVs of parents tested in each region have been estimated by BLUP based on historical family trials over about 10 years in the corresponding region. This paper examined these EBVs for the parents tested in two series of family trials (stage 1 in 1995 and 1996 series) and the two subsequent series of clonal testing trials (stage 3 in 2000 and 2001 series) from Central and Burdekin regions to link the EBVs of parents in stage 1 and those in stage 3 and then to investigate the correlations between EBVs of parents and their progeny performance or clonal mean. Crop traits investigated were commercial cane sugar (CCS), tonnes of cane per hectare (TCH) and net merit grade (NMG, an index of CCS, tonnage of sugar and fibre content, relative to standards; the average NMG of standards in a trial is constrained to 10). The correlations between EBVs of parents and clonal means were low and generally insignificant, 0.15 to 0.24 for CCS, 0.06 to 0.18 for TCH and –0.06 to 0.18 for NMG. Some higher EBV parents failed to produce any clones in stage 3 while poor EBV parents had progeny represented. These unexpected results could be caused by the poor estimation of EBVs and the inherent poor relationship between parental EBV and clonal performance. Approaches to improve the estimation of EBV and to address the poor relationship are proposed and discussed.