| Summary: | Accurate prediction of flowering time helps breeders to develop new varieties that can achieve maximal efficiency in a changing climate. A methodology was developed for the construction of a simulation model for flowering time in which a function for daily progression of the plant from one to the next phenological phase is obtained in analytic form by stochastic minimization. The resulting model demonstrated high accuracy on the recently assembled data set of wild chickpeas. The inclusion of genotype-by-climatic factors interactions accounted to 77% of accuracy in terms of root mean square error. It was found that the impact of minimal temperature is positively correlated with the longitude at primary collection sites, while the impact of day length is negatively correlated. It was interpreted as adaptation of accessions from highlands to lower temperatures and those from lower elevation river valleys to shorter days. We used bootstrap resampling to construct an ensemble of models, taking into account the influence of genotype-by-climatic factors interactions and applied it to forecast the time to flowering for the years 2021–2099, using generated daily weather in Turkey, and for different climate change scenarios. Although there are common trends in the forecasts, some genotypes and SNP groups have distinct trajectories.
|
|---|
