Initiation of genomic recurrent selection for slow-rusting adult plant resistance to stem rust in wheat
Rutkoski, J.
Initiation of genomic recurrent selection for slow-rusting adult plant resistance to stem rust in wheat - 2012 - p. 167
Abstract only
Genomic selection (GS) is a new breeding technology that promises to increase the rate of genetic gain for quantitative traits. With GS, a prediction model for the trait(s) of interest is developed using a relevant population with existing phenotypic and genotypic data. The prediction model is then applied to a new set of breeding lines which have been genotyped in order to estimate the breeding values. Based on the predicted breeding values, lines can be selected for advancement or crossing to initiate the next selection cycle. This allows selection to|occur before phenotypic information for the lines can be generated, therefore enabling faster selection cycles. In order to empirically evaluate the efficiency of GS relative to phenotypic selection (PS) for slow-rusting adult plant stem rust resistance (APR), two parallel recurrent selection (RS) schemes are underway, and are currently at the first selection cycle stage.Selections are based on phenotypes in the first scheme (RS-PS), and on genotypes and predicted breeding values in the second scheme (RS-GS). Based on cross-validation using the model training population, we expect the selection accuracy with GS to be 0.6. Based on phenotypic data from the same germplasm evaluated across four growing seasons, we expect the selection accuracy with PS to be 0.78. Because the minimum selection cycle duration is 8 months with GS and 12 months with PS, considering our estimates of selection accuracy we expect to achieve 1.15 times more genetic gain per unit of time with GS compared to PS.
English
978-0-615-70429-6
Initiation of genomic recurrent selection for slow-rusting adult plant resistance to stem rust in wheat - 2012 - p. 167
Abstract only
Genomic selection (GS) is a new breeding technology that promises to increase the rate of genetic gain for quantitative traits. With GS, a prediction model for the trait(s) of interest is developed using a relevant population with existing phenotypic and genotypic data. The prediction model is then applied to a new set of breeding lines which have been genotyped in order to estimate the breeding values. Based on the predicted breeding values, lines can be selected for advancement or crossing to initiate the next selection cycle. This allows selection to|occur before phenotypic information for the lines can be generated, therefore enabling faster selection cycles. In order to empirically evaluate the efficiency of GS relative to phenotypic selection (PS) for slow-rusting adult plant stem rust resistance (APR), two parallel recurrent selection (RS) schemes are underway, and are currently at the first selection cycle stage.Selections are based on phenotypes in the first scheme (RS-PS), and on genotypes and predicted breeding values in the second scheme (RS-GS). Based on cross-validation using the model training population, we expect the selection accuracy with GS to be 0.6. Based on phenotypic data from the same germplasm evaluated across four growing seasons, we expect the selection accuracy with PS to be 0.78. Because the minimum selection cycle duration is 8 months with GS and 12 months with PS, considering our estimates of selection accuracy we expect to achieve 1.15 times more genetic gain per unit of time with GS compared to PS.
English
978-0-615-70429-6