Chapter 13 - Gene-Based Crop Models

Abstract

The purpose of this chapter is to explain how one can incorporate genetic information into crop models for predicting the phenotype from information on the genotype and environment. Previously, we have assumed, explicitly or implicitly, that a crop model describes the behavior of a single crop genotype as a function of environment and management. The explanatory variables have been environmental variables (e.g., temperature) and management variables (e.g., planting date). In this chapter, we consider models that describe the behavior of a population of genotypes as well as behavior of individual genotypes contained in that population. Explanatory variables are environmental variables and management variables as before, but now also include variables that describe the genetic make-up of each genotype. In this chapter, we define basic concepts, describe data that are needed to develop and evaluate gene-based models, describe a population of the common bean (Phaseolus vulgaris L.) created from two parents of very different genetic backgrounds, and show methods for developing and evaluating the ability of a gene-based model to predict time to first flower for genotypes in this population. The methods address the following questions: What genetic variables are needed, how to insert the genetic variables into the model, and how to estimate parameters for multiple genotypes grown in multiple environments. Empirical results are shown for the dynamic gene-based model and for a statistical model, both developed to predict time to first flower. The dynamic model described 88% of the overall variability in the observations on time to first flower, 81% of the variation in time to first flower caused by differences among genotypes, 98% of the variation caused by variations in environmental variables, and about 47% of the variation associated with interactions between G and E (GEI). The approaches presented are based on a biparental population, but describe concepts that can be used to develop more comprehensive dynamic models for more diverse populations and environments.