Simulation-Based Analysis of Effects of Vrn and Ppd Loci on Flowering in Wheat

Authors: White, Jeffrey; HERNDL, MARKUS - UOFHOHENHEIM,GERMANY; HUNT, L - UOFGUELPH,ONTARIO CANADA; PAYNE, THOMAS - CIMMYT, MEXICO; HOOGENBOOM, GERRIT - UOFGA, GRIFFIN GEORGIA

Submitted to: Crop Science
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 9/20/2007
Publication Date: 3/1/2008
Citation: White, J.W., Herndl, M., Hunt, L.A., Payne, T.S., Hoogenboom, G. 2008. Simulation-Based Analysis of Effects of Vrn and Ppd Loci on Flowering in Wheat. Crop Science, (48):678-687.

Interpretive Summary: The time of heading or flowering in wheat can strongly influence grain yield and quality. Wheat simulation models predict heading date based on information on daily weather, location, and cultivar type. Cultivar differences are accounted for through model parameters that adjust how low temperatures (through a response called vernalization) and daylength affect wheat development. These parameters are usually estimated by adjusting their values until comparisons of simulated and observed data reach satisfactory agreement. This trial and error process is inefficient in that it usually requires multiple seasons of data and that errors in calibration can hide problems with other data or the model itself. Work with common bean and soybean showed that it was possible to estimate cultivar parameters based on information about actual genes present in different cultivars. For bread wheat, available data on genes affecting vernalization and photoperiod, the Vrn and Ppd families of genes, appeared sufficient to estimate two parameters in the model CSM-Cropsim-CERES. The parameters, which affect vernalization and photoperiodism, were first estimated using the conventional procedures and then re-estimated using linear regressions for effects of the Vrn and Ppd loci. Flowering data were obtained for 29 cultivars tested in the International Winter Wheat Performance Nursery, which was distributed from 1968 to 1981 by the University of Nebraska with funding from USDA and USAID. For our analyses, the experimental data were divided into geographically distinct sets, one for calibration (14 locations, 540 observations) and one evaluation (34 locations, 1499 observations). Using gene-based estimates of two parameters that determine the vernalization requirement and photoperiodism in the model, 96% of variation in flowering was explained for the calibration dataset and 90% for the evaluation set, with average errors of 9 days and 10 days for the two datasets, respectively. The conventional coefficients explained an additional 1% of variation for each dataset. Recent advances in characterizing the wheat genome give promise for enabling rapid, reliable identification of genes in specific cultivars. Gene-based prediction of phenology thus appears feasible for improving predictions of how wheat genotypes respond to environment, allowing more accurate targeting of cultivars to regions or production conditions and for examining issues such as crop response to climate risk or global warming.

Technical Abstract: Grain yield and quality in cereals are often strongly influenced by flowering date. Ecophysiological models of bread wheat (Triticum aestivum L.) simulate the number of days to heading or anthesis by assuming that an intrinsic rate of development is modified by vernalization and photoperiodism. Cultivar differences are accounted for through parameters for vernalization requirement, photoperiod sensitivity, and earliness per se. These parameters are usually estimated by optimization based on comparison of simulated and observed data on development. In other crops, similar parameters have been estimated based on the genetic makeup of individual cultivars. For bread wheat, available data on the Vrn and Ppd loci, which affect vernalization and photoperiodism, appeared sufficient to estimate two parameters in the CSM-Cropsim-CERES using cultivar haplotypes. The parameters, which affect vernalization and photoperiodism, were first estimated using conventional procedures and then re-estimated using linear regressions for effects of the Vrn and Ppd loci. Flowering data were obtained for 29 cultivars tested in the International Winter Wheat Performance Nursery, which was distributed from 1968 to 1981 by the University of Nebraska. For our analyses, the experimental data were divided into geographically distinct sets, one for calibration (14 locations, 540 observations) and one evaluation (34 locations, 1499 observations). Using gene-based estimates of two parameters determining vernalization requirement and photoperiodism in the model, 96% of the variation in flowering was explained for the calibration dataset and 90% for the evaluation set, with root mean squared errors of 9 days and 10 days for the two datasets, respectively. The conventional coefficients explained an additional 1% of variation for each dataset. Recent advances in characterizing the wheat genome give promise for enabling rapid, reliable determination of haplotypes. Gene-based prediction of phenology appears feasible for improving predictions of how wheat genotypes respond to environment, allowing more accurate targeting of germplasm and analysis of issues such as crop response to climate risk or global warming.