Gene-based approaches to crop simulation: Past experiences and future opportunities

Use of process-based models of plant growth and development is increasing in both basic and applied research. Advances in genomics suggest the possibility of using information on gene action to improve simulation models, particularly where differences among genotypes are of interest. This paper reviews issues related to incorporating gene action in crop models, starting with an introduction to basic concepts of functional genomics. We recognize six levels of genetic detail in modeling approaches. Modeling gene action through linear estimates of effects on model parameters (Level 4) has shown promise in the common bean (Phaseolus vulgars L.) model GeneGro. However, this approach requires extensive data on the genetic makeup of cultivars, and such data are still not routinely available. Software for simulating complex biochemical pathways offers the prospect of simulating processes such as photosynthesis or photoperiod control of flowering by considering interactions of regulators, gene-products, and other metabolites (Level 6), but such software applications may require an understanding of the reaction kinetics of large biomolecules existing at concentrations as low as one or two molecules per cell. Over the next decade, genetic information probably has the most to contribute in understanding temporal and tissue-level variation in the genetic control of specific processes and, for more applied modeling, in improving the representation of cultivar differences. Strategic decisions are needed on prioritization among species and traits to be modeled, as well as on how to improve collaboration with molecular biologists to better access and harness the data resulting from their research.