SIMULATING EFFECTS OF GENES FOR PHYSIOLOGICAL TRAITS IN A PROCESS-ORIENTED CROP MODEL

Recent improvements in crop simulation techniques and in understanding of crop genetics suggest the possibility of integrating genetic infor mation on physiological traits into crop simulation models. By using k nown genotypes, rather than empirically fitted cultivar-specific coeff cients, a simulation model should permit more explicit, testing of hyp otheses concerning the genetic basis of adaptation of cultivars to dif ferent environments or production systems. This paper describes and ev aluates GeneGro, a version of the dry bean (Phaseolus vulgaris L.) cro p simulation model BEANGRO version 1.01 modified to incorporate effect s of seven genes affecting phenology: growth habit, and seed size: Ppd , Hr, Fin, Fd, and Ssz-1, and two more genes for seed size inferred fr om indirect evidence. Thirty cultivars were calibrated for BEANGRO usi ng data from 14 trials conducted in Colombia, Guatemala, Mexico, and F lorida. The resulting cultivar-specific coefficients of BEANGRO were r eplaced with information on specific genotypes of cultivars to create the gene-based model. With cultivar differences specified only by the seven genes, GeneGro explained 31% of observed variation for seed yiel d, 58% for seed weight, 84% for days to flowering, 85% for days to mat urity 52% for maximum leaf area index, and 36% for canopy dry weight a t maturity, but 0% for harvest index. In testing the effectiveness of GeneGro after overall trial and cultivar effects were accounted for th rough regression analysis, all simulated data except for seed weight s howed significant relations with observed data (P less than or equal t o 0.01). Our results indicate that for certain traits surprisingly few genes must be characterized to simulate cultivar differences as accur ately as with the BEANGRO model. Furthermore, they suggest a potential for developing models similar to GeneGro for studying the effects of genes on adaptation in other crops.