Submitted to: Illinois Corn Breeders School Proceedings
Publication Type: Proceedings
Publication Acceptance Date: February 3, 2007
Publication Date: March 6, 2007
Citation: Edwards, J.W. 2007. Estimating Genotype- and Environment-Specific Heritabilities. Illinois Corn Breeders School Proceedings. p. 228-241. Interpretive Summary: Identifying plant varieties (commercial and experimental) that provide superior performance for growers across a wide range of growing conditions is a very expensive task. Varieties must be evaluated in many environments and performance data summarized in order to identify not only the best varieties, but also those with the most stable performance. A new method for more efficiently summarizing the data from plant variety evaluations has been tested and will be described in this talk. The method increases precision of plant variety evaluations because greater value will be realized from existing data, and will potentially reduce cost as the number of locations and replications in such evaluations may be reduced. As such, use of this method will benefit all parties involved in plant variety evaluation including public and private plant breeders, seed companies, and public yield testing programs. Growers will realize indirect benefits in the form of improved varieties, more stable varieties in particular. Growers will benefit directly from better information available on existing commercial and public varieties.
Technical Abstract: The advantages of computing genotype- and environment-specific heritabilities are discussed. A statistical approach is used in which logvariances of both genotype by environment interaction and error are modeled as random variables. Resulting estimators of variances are weighted averages of a pooled variance and a cultivar- or genoytype-specific variance. It is shown that in the presence of heterogeneous variances, shruken estimators of variances are more precise (lower mean-squared error) than either pooled variance estimators or individual genotype or environment estimators. The implications in terms of precision of estimation of performance and ranking of performance are discussed. Models with heterogeneous variances have slightly lower means-squared errors, on average, of cultivar performance estiamtors than traditional blup. However, a greater advantage may be seen when comparing cultivar rankings because unstable cutlivars have higher estimated gneoytpe by environment interactions, which is reflected with lower precision estimators for unstable cultivars and performance estimators that are shrunken towards the mean more than for stable cultivars.