A further modification in the triple testcross analysis when testers share common loci

Abstract

For investigating the genetical structure of a population of homozygous lines, modifications in the triple testcross analysis that allow for the effect of common loci shared by the pure-breeding testers, L1 and L2, have been previously proposed. The unbiased estimates of the additive genetic and dominance components are, however, correlated as they are obtained from statistics that are nonorthogonal. Further modifications are suggested where full triple testcross families L1i (= Pi × L1), L2i (= Pi × L2) and L3i (= Pi × L3; L3 = L1 × L2) along with selfed progenies of L1i (= L1si) and L2i (= L2si) and pure-breeding lines (Pi) themselves are raised. Alternatively, full triple testcross progenies (L1i, L2i, and L3i) can be raised along with the selfs of L3i (= L3si) families and parental (Pi) lines. These combinations of normal and selfed progenies allow one to test and estimate unbiased additive and dominance components, free from the effect of common loci in the testers and from statistics that are statistically orthogonal. The components of variation are, therefore, estimated independently.Key words: triple testcross, inadequate testers, components of variation.