Optimal Mating of Pinus taeda L. Under Different Scenarios Using Differential Evolution Algorithm

Abstract

Abstract A newly developed software, AgMate, was used to perform optimized mating for monoecious Pinus taeda L. breeding. Using a computational optimization procedure called differential evolution, AgMate was applied under different breeding population sizes scenarios (50, 100, 150, 200, and 250) and candidate contribution scenarios (maximum use of each candidate was set to 1 or 8), to assess its efficiency in maximizing the genetic gain while controlling inbreeding. A population of 962 Pinus taeda parents with a known pedigree from the North Carolina State University Tree Improvement Program was used to optimize objective functions accounting for the coancestry of parents and expected genetic gain and inbreeding of the future progeny. AgMate results were compared with those from another widely used mating software called MateSel. For the proposed mating list of 200 progenies, AgMate resulted in an 83.7% increase in genetic gain compared with the candidate population. There was evidence that AgMate performed similarly to MateSel in managing coancestry and expected genetic gain, but MateSel was superior in avoiding inbreeding in proposed mate pairs. The developed algorithm was computationally efficient in maximizing the objective functions and flexible for practical application in monoecious diploid conifer breeding. AgMate, with its open-source software, free-to-modify algorithm and front-end ShinyApp, is a necessary addition for the advancement of conifer breeding. Study Implications: A dataset from a breeding population of loblolly pine (Pinus taeda L.) was analyzed using an optimal mating software, AgMate (developed by the authors), to optimize the selection, contribution, and mating of candidates simultaneously. The software helps breeders decide on trees to cross and the crossing frequency, such that the trees are unrelated and would result in the best-performing progenies. AgMate is effective in meeting the breeding objectives for monoecious diploid species. The open-source, easy-to-use, and flexible AgMate software, also accessible via a website, is invaluable in helping breeders create optimal matings for future generations, which balances the pursuit of maximizing genetic gain while maintaining genetic diversity.