Mating structures for genomic selection breeding programs in aquaculture.

Anna K Sonesson,Jørgen Ødegård

doi:10.1186/s12711-016-0224-y

Anna K Sonesson, Jørgen Ødegård

Open Access

https://doi.org/10.1186/s12711-016-0224-y

Copy DOI

Abstract

BackgroundIn traditional family-based aquaculture breeding, each sire is mated to two dams in order to separate the sire’s genetic effect from other family effects. Factorial mating involves more mates per sire and/or dam and result in more but smaller full- and/or half-sib families. For traits measured on sibs of selection candidates, factorial mating increases intensity of selection between families when selection is on traditional best linear unbiased prediction (BLUP) estimated breeding values (TRAD-EBV). However, selection on genome-wide estimated breeding values (GW-EBV), uses both within- and between-family effects and the advantage of factorial mating is less obvious. Our aim was to compare by computer simulation the impact of various factorial mating strategies for truncation selection on TRAD-EBV versus GW-EBV on rates of inbreeding, accuracy of selection and genetic gain for two traits, i.e. one measured on selection candidates (CAND-TRAIT) and one on their sibs (SIB-TRAIT).ResultsSire:dam mating ratios of 1:1, 2:2 or 10:10 were tested with 100, 200 or 1000 families produced from a constant number of parents (100 sires and 100 dams), and a mating ratio of 1:2 with 200 families produced from 100 sires and 200 dams. With GW-EBV, changing the mating ratio from 1:1 to 10:10 had a limited effect on genetic gain (less than 5 %) for both CAND-TRAIT and SIB-TRAIT, whereas with TRAD-EBV, selection intensity increased for SIB-TRAIT and genetic gain increased by 41 and 77 % for schemes with 3000 and 12,000 selection candidates, respectively. For both GW-EBV and TRAD-EBV, rates of inbreeding decreased by up to ~30 % when the mating ratio was changed from 1:1 to 10:10 for schemes with 3000 to 12,000 selection candidates. Similar results were found for alternative heritabilities of SIB-TRAIT and total number of tested sibs.ConclusionsChanging the sire:dam mating ratio from 1:1 to 10:10 increased genetic gain substantially with TRAD-EBV, mainly through increased selection intensity for the SIB-TRAIT, whereas with GW-EBV, it had a limited effect on genetic gain for both traits. Rates of inbreeding decreased for both selection methods.

Highlights

In traditional family-based aquaculture breeding, each sire is mated to two dams in order to separate the sire’s genetic effect from other family effects
For traits that cannot be measured on selection candidates, but that are instead measured on sibs of candidates, traditional best linear unbiased prediction (BLUP) assigns identical breeding value predictions to all nonphenotyped members of a family
With traditional BLUP estimated breeding values, factorial mating is expected to increase the intensity of selection between families on traits that are measured on sibs of the candidates

Summary

Introduction

In traditional family-based aquaculture breeding, each sire is mated to two dams in order to separate the sire’s genetic effect from other family effects. For traits measured on sibs of selection candidates, factorial mating increases intensity of selection between families when selection is on traditional best linear unbiased prediction (BLUP) estimated breeding values (TRAD-EBV). With traditional BLUP estimated breeding values, factorial mating is expected to increase the intensity of selection between families on traits that are measured on sibs of the candidates (since more families are available for the same number of parents). For genome-wide estimated breeding values, even for traits that are measured on sibs of the candidates, selection is done both within and between families and the advantage of factorial mating is less obvious

Objectives

Methods

Results

Discussion

Conclusion