Genetic parameter estimates for resistance to rust (Cronartium quercuum infection from full-sib tests of slash pine (Pinus elliottii), modelled as functions of rust incidence

Data from 171 slash pine progeny tests, incorporating over 700 different families from more than 2100 first-generation parents and approximately 170000 trees, were used to estimate variance and covariance components by Restricted Maximum Likelihood (REML) in both single-site and paired-site analyses. From these REML estimates, genetic parameters (heritabilities, proportion of dominance, type B genetic correlations, and age-age genetic correlations) were estimated for resistance to fusiform rust infection at 4 to 15 years of age. Predictive models were developed for biased (single-site) heritability,unbiased (paired-site) heritability and the type B genetic correlation. Biased heritability exhibited a maximum of 0.20 at an average rust infection of 72%. Unbiased heritability estimates from paired-site analyses increased linearly with increasing average rust infection in the tests; however, in very few test pairs did the average rust infection exceed 75%, andparents and approximately 170000 trees, were used to estimate variance and covariance components by Restricted Maximum Likelihood (REML) in both single-site and paired-site analyses. From these REML estimates, genetic parameters (heritabilities proportion of dominance, type B genetic correlations, and age-age genetic correlations) were estimated for resistance to fusiform rust infection at 4 to 15 years of age. Predictive models were developed for biased (single-site) heritability, unbiased (paired-site) heritability and the type B genetic correlation. Biased heritability exhibited a maximum of 0.20 at an average rust infection of 72%. Unbiased heritability estimates from paired-site analyses increased linearly with increasing average rust infection in the tests; however, in very few test pairs did the average rust infection exceed 75%, and extrapolation beyond 67% would be unreliable. Although genotype-environment interaction was present, it was of little consequence except at low rust infection levels. The proportion of dominance variance (when compared to the total phenotypic variance) was not related to rust infection levels or age, and across all tests and ages averaged 0.087 and 0.053 in singlesite and paired-site analyses respectively. Since dominance variance was small relative to additive variance, except when rust infection levels were low, it was considered to be of little practical importance. Test age was not a significant factor affecting any genetic parameter examined, and all age-age genetic correlations were near +1.0. The results endorse the current breeding strategy, which is based on recurrent selection for general combining effects, and stresses the importance of restricting improvement efforts to sites with moderate to high rust infection levels