Comparison of efficiency of F1 and F2 generations in identifying potential inbreds for improvement of a reference maize hybrid

Abstract

The estimators namely, the relative frequency of favourable dominant alleles in a donor inbred (/lG) , the net improvement (NI) and the probability of net gain of favourable dominant alleles (PNGg) were compared for their efficiency in ranking donors both within and between two methods, one using inbred and F, data and the other using inbred and F2 data. Six donor maize inbreds (TCA21, TCA22, TCA24, CML32, IPA40 and SC7) were crossed to each of the parents of a reference maize hybrid, IPA2 x IPA34. Thirteen F,'s, 13 F2's and 8 inbreds were evaluated in two locations in 1996 for grain yield. Inbreds with the highest and the lowest estimates of /lG, NI and PNGg were the same in both the methods. The efficiency of different estimators in identifying donors was highly consistent within a method and comparison of efficiency among them did not change with the method used. The ranking of donors as per /lG and PNGg varied greatly, while that as per NI remained the same with the method used. The donors with the closest relative relationship to either parent of the reference hybrid were the same in both the data sets. The differences in ranking of donors, different assessments of relative relationships of lines and different decisions on where to backcross prior to selfing may have stemmed from failure of assumptions such as no epistasis or difference in precision in the evaluation of different generations. The study indicated that direct selfing of the crosses IPA2 x TCA22 and IPA34 x SC7 would release an improved version of IPA2 and IPA34, respectively in a relatively shorter period.