Breeding objectives aim to optimize two crucial but contrasting goals of maximizing genetic gain while managing genetic diversity. In advanced generations, this becomes a challenge in monoecious conifer tree species breeding programs because they suffer from inbreeding. Developing an algorithm that maximizes genetic gain while maintaining genetic diversity for monoecious species is imperative. While methods and algorithms for animal breeding are well-established, an efficient algorithm suited to monoecious species remains elusive. Towards this goal, we have adopted an evolutionary genetic algorithm, the Differential Evolution algorithm, to optimize mate pair designing in Pinus taeda (loblolly pine), a widely planted pine species in the southern USA. AgMate, an optimal mating for monoecious species software, is a multi-functional, completely automated optimization software. It utilizes genetic relationships and breeding values as input to create an optimal mating list. AgMate maximizes the genetic gain and minimizes the increase in average coancestry and inbreeding in the proposed progeny. AgMate was more effective in optimizing mating lists than positive assortative mating and random mating in short-term and long-term settings. AgMate mating list resulted in an average 93% genetic gain each cycle for ten cycles while simultaneously minimizing the increase in coancestry to 0.086. The framework and methods adapted for Pinus taeda are also relevant to the breeding of other monoecious species.

Agmate, an R software that utilizes the Differential Evolution algorithm and Second-order Cone Programming for the optimization of mate pair decisions and contributions for monoecious diploid species like Pinus taeda was used to analyze the data. 

AgMate is a multi-functional software able to

Process pedigree information

1. Complete and order pedigrees

2. Generate numerator relationship matrices

3. Optimize selection and contribution of candidates

4. Design optimal matings between candidates

5. Provide visualization of results

Detailed R code is available at https://github.com/khushigoda/AgMate.

The ShinyApp version is available here: https://khushigoda.shinyapps.io/AgMate/.