Robust designs involving partial triallel crosses for breeding experiments

Abstract

Obtaining information regarding general combining ability (gca) and specific combining ability (sca) effects can be the major objective of a breeding programme to release new hybrids with enhanced genetic potential. This study included designs involving partial triallel crosses as they, being genetically more viable and consistent performers, involve lesser number of crosses leading to a lower degree of fractionation. An optimal or efficient design for triallel cross experiment may become disconnected and inefficient if the underlying assumptions are not fulfilled due to a missing observation pertaining to a cross. The robustness of designs against missing observation using connectedness and efficiency criteria has been studied both under unblocked and blocked situations. A list of efficient robust designs for triallel cross experiments has been tabulated for both unblocked and blocked situation. Programs have been written in SAS [PROC IML] software for computing efficiency factor of the designs involving triallel crosses for estimating Gca effects to investigate the robustness of designs against missing observation by calculating the efficiency factor.