The Modular Breeding Program Simulator (MoBPS) allows efficient simulation of complex breeding programs

Abstract

Context.Breeding programs aim at improving the genetic characteristics of livestock populations with respect to productivity, fitness and adaptation, while controlling negative effects such as inbreeding or health and welfare issues. As breeding is affected by a variety of interdependent factors, the analysis of the effect of certain breeding actions and the optimisation of a breeding program are highly complex tasks.Aims. This study was conducted to display the potential of using stochastic simulation to analyse, evaluate and compare breeding programs and to show how the Modular Breeding Program Simulator (MoBPS) simulation framework can further enhance this.Methods. In this study, a simplified version of the breeding program of Göttingen Minipigs was simulated to analyse the impact of genotyping and optimum contribution selection in regard to both genetic gain and diversity. The software MoBPS was used as the backend simulation software and was extended to allow for a more realistic modelling of pig breeding programs. Among others, extensions include the simulation of phenotypes with discrete observations (e.g. teat count), variable litter sizes, and a breeding value estimation in the associated R-package miraculix that utilises a graphics processing unit.Key results. Genotyping with the subsequent use of genomic best linear unbiased prediction (GBLUP) led to substantial increases in genetic gain (15.3%) compared with a pedigree-based BLUP, while reducing the increase of inbreeding by 24.8%. The additional use of optimum genetic selection was shown to be favourable compared with the plain selection of top boars. The use of graphics processing unit-based breeding value estimation with known heritability was~100 times faster than the state-of-the-art R-package rrBLUP.Conclusions. The results regarding the effect of both genotyping and optimal contribution selection are in line with well established results. Paired with additional new features such as the modelling of discrete phenotypes and adaptable litter sizes, this confirms MoBPS to be a unique tool for the realistic modelling of modern breeding programs.Implications. The MoBPS framework provides a powerful tool for scientists and breeders to perform stochastic simulations to optimise the practical design of modern breeding programs to secure standardised breeding of highquality animals and answer associated research questions.

Beyond Scenarios - Optimization of breeding program design (MoBPSopti)

“…Combined with the high number of parameters to consider, it becomes computationally challenging to simulate each potential breeding scheme and directly derive the optimal one. Therefore, to overcome these computational challenges, analysis of breeding programs via stochastic simulations is usually limited to a couple of potentially interesting scenarios, which are then simulated and compared against each other (Wensch-Dorendorf et al 2011;Esfandyari et al 2015;Gaynor et al 2017;Büttgen et al 2020;Pook et al 2021).…”

“…The grid-search approach is a commonly used algorithm for finding the optimum allocation of test resources when there are few parameters to optimize (Longin et al 2006;Gordillo and Geiger 2008;Mi et al 2014Mi et al , 2016Pook et al 2021). However, although grid-search provides an acceptable solution to the problem in many smaller applications, as it is relatively straightforward to define a grid of possible parameter combinations and evaluate the performance of each combination, it is not efficient for a large number of parameters due to the increased computational time and, or the need for a sparse grid.…”