Simulation of the comparative gains from four different hybrid tree breeding strategies

Abstract

There is increasing interest in the deployment of interspecific hybrids in forest tree planting. The associated breeding programs are usually an adaptation of the reciprocal recurrent selection (RRS) strategy outlined by Comstock et al. (R. Comstock, H. Robonson, and P. Harvey. 1949. Agron. J. 41: 360–367) or use recurrent selection for general combining ability (GCA) in the pure species. This study uses a computer simulation tool known as XSIM, which has been described in a previous paper, to investigate the efficiency of four hybrid strategies. In addition to conventional RRS, we considered RRS with forward selection (RRS-SF), a strategy that approximately halves the generation interval needed for RRS, because hybrid and pure species progeny are bred simultaneously. Forward and backward selections are also made simultaneously and not in successive generations as is the case for RRS. An innovative best linear unbiased prediction analysis makes this possible. The development of a synthetic species (SYN) and pure species selection (PSS) were other strategies tested. The strategies were tested across a wide range of genetic structures. Genetic structures were defined as particular combinations of the correlation between pure species and hybrid performance for each species and the proportion of the genetic variance that is additive, dominance, and epistatic for each species. The results of the simulation have shown that the SYN strategy is the most cost effective across a wider range of genetic structures. This is especially so for those structures where there is less dominance variance and the pure–hybrid correlations in both species are greater than zero. Where the SYN strategy is not cost effective, the RRS-SF strategy is then the best option.