"Selection Strategies for Tradeoffs in Recurrent Genomic Selection in Simulated SoyNAM Populations."

Project Description: Management, genetic and developmental constraints impose limits on crop improvement. The discovery and implementation of optimal means to deal with genetic constraints is an important aspect of long-term crop improvement as yield of many crops have reached stagnation much below their potential, which affects global output of these crops. Genetic constraints in crop improvement systems include the depletion of genetic variance in a population and the loss of rare favorable alleles in early cycles of selection, which affect the realization of potential in later cycles. In this study, we have compared three approaches that include weighted genomic selection, implementations of parallel and multi-objective genetic algorithm (GA) based methods for optimal long-term genetic gain in recurrent genomic selection using simulated SoyNAM system. We have also investigated hybrid approaches that combine elements of these strategies to deal with tradeoffs for long-term genetic gain. We have used factorial design and non-linear modelling to investigate the impact of 2664 unique combinations of factors on response to selection. Results from this study have the potential to help plant breeders make decisions regarding choice of crosses for optimal long-term genetic improvement of crops and consider new approaches that combine the power of prediction modeling, optimization methods and knowledge of genetics for dealing with tradeoffs in long-term genetic improvement of crops.

Supplementary resources including data and R package for Vishnu's work: "Selection Strategies for Tradeoffs in Long-term Recurrent Genomic Selection"

Documentation for 'SoyNAMSelectionMethods' R Package

  • Methods for Non-linear Models -Part I Process for fitting non-linear multi-level random intercept model using 'nlme' R package

  • Methods for Non-linear Models -Part II Process for fitting single-level random intercept non-linear models using 'nlme' R package for cycle subsets

  • Methods for Non-linear Models -Part III M15 fit with correlation structure

  • Evaluation Metrics Description of evaluation metrics used in this study

  • R package and Data

  • Download SoyNAMSelectionMethods R package "SoyNAMSelectionMethods_0.1.0.tar.gz"