Abstract
The next generation of gene-based crop models offers the potential of predicting crop vegetative and reproductive development based on genotype and weather data as inputs. Here, we illustrate an approach for developing a dynamic modular gene-based model to simulate changes in main stem node numbers, time to first anthesis, and final node number on the main stem of common bean (Phaseolus vulgaris L.). In the modules, these crop characteristics are functions of relevant genes (quantitative trait loci (QTL)), the environment (E), and QTL×E interactions. The model was based on data from 187 recombinant inbred (RI) genotypes and the two parents grown at five sites (Citra, FL; Palmira, Colombia; Popayan, Colombia; Isabela Puerto Rico; and Prosper, North Dakota). The model consists of three dynamic QTL effect models for node addition rate (NAR, No. d−1), daily rate of progress from emergence toward flowering (RF), and daily maximum main stem node number (MSNODmax), that were integrated to simulate main stem node number vs. time, and date of first flower using daily time steps. Model evaluation with genotypes not used in model development showed reliable predictions across all sites for time to first anthesis (R2=0.75) and main stem node numbers during the linear phase of node addition (R2=0.93), while prediction of the final main stem node number was less reliable (R2=0.27). The use of mixed-effects models to analyze multi-environment data from a wide range of genotypes holds considerable promise for assisting development of dynamic QTL effect models capable of simulating vegetative and reproductive development.
Highlights
Tools that integrate genetic, environment and management information to predict crop performance in contrasting environments are needed to meet global food demands and assist plant breeders in designing new cultivars for increased yield (Hatfield and Walthall, 2015)
quantitative trait loci (QTL) analyses can dissect the genetic architecture of complex traits, and in combination with statistical methods, such as mixed effect models, it is possible to estimate the genetic, environmental, and G × E effects on the phenotype (Boer et al, 2007; Chenu et al, 2009; Peiffer et al, 2014). We propose that these mixed effect approaches can be used to identify QTL, E, and QTL × E interactions underlying specific crop processes and that together with the decades of understanding of processes mechanisms from crops models can be combined to build a gene-based crop model that predicts aspects of crop performance based on genetic, environment, and management data
Given the fact that recombination is significantly suppressed in this region, it is highly unlikely that these QTLs will be resolved by recombinational analysis
Summary
Environment and management information to predict crop performance in contrasting environments are needed to meet global food demands and assist plant breeders in designing new cultivars for increased yield (Hatfield and Walthall, 2015). Of note is that parameters termed “genetic coefficients” or “Genotype-Specific Parameters (GSPs)” that describe phenology, plant architecture (leaf area, number and plant dimensions), and biomass allocation in existing crop models are not yet linked to any gene(s) They do not take into account gene-by-environment (G × E) or G × G interactions at the level of individual processes that are considered in the models. This lack of genetic information within the crop models requires multi-environment experiments to estimate the GSP values when new cultivars (genotypes) are released This process is time consuming, costly, and limits the utility of crop models in plant breeding programs and other practical applications. A step is to integrate genetic information (G and G × E) into models to predict a genotype's performance in a targeted environment
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
