Simulated Individual Best Linear Unbiased Prediction versus Mass Selection in Sugarcane Families.

The purpose of this study was to compare the method BLUPIS (best linear unbiased prediction individual simulated) with mass selection in terms of efficiency in identifying the best genotypes in sugarcane (Saccharum officinarum L.) families. Mass selection was performed by two breeders with 25 yr of experience. The BLUPIS procedure selected families with higher means for tons of cane per hectare (TCH) than the over-all mean. The number of plants selected per family was calculated by n_k = (ĝ_k / ĝ_j )n_j, where ĝ_k indicates the genotypic value of the kth family; ĝ_j the genotypic value of the best family; and n_j is equal to the number of plants selected in the best family, determined as n_j = 45 in this study. Out of 20 best clones forwarded to the third test phase (T3), BLUPIS selected all in the first test phase (T1) and mass selection only two. Therefore, 100% of the clones in the second test phase (T2) had been selected by BLUPIS. The BLUPIS was most efficient in detecting the best genotypes, since all clones that were promoted up to phase T3 were descendants from the best families. The BLUPIS method should be applied in sugarcane breeding programs to ensure the selection of the best genotypes. [ABSTRACT FROM AUTHOR]