Introgression of a major QTL from an inferior into a superior population using genomic selection

Abstract

Background: Selection schemes aiming at introgressing genetic material from a donor into a recipient line may be performed by backcross-breeding programs combined with selection to preserve the favourable characteristics of the donor population. This stochastic simulation study investigated whether genomic selection can be effective in preserving a major quantitative trait locus (QTL) allele from a donor line during the backcrossing phase.

Methods: In a simulation study, two fish populations were generated: a recipient line selected for a production trait and a donor line characterized by an enhanced level of disease resistance. Both traits were polygenic, but one major QTL affecting disease resistance was segregating only within the donor line. Backcrossing was combined with three types of selection (for total merit index) among the crossbred individuals: classical selection, genomic selection using genome-wide dense marker maps, and gene-assisted genomic selection. It was assumed that production could be observed directly on the selection candidates, while disease resistance had to be inferred from tested sibs of the selection candidates.

Results: Classical selection was inefficient in preserving the target QTL through the backcrossing phase. In contrast, genomic selection (without specific knowledge of the target QTL) was usually effective in preserving the target QTL, and had higher genetic response to selection, especially for disease resistance. Compared with pure genomic selection, gene-assisted selection had an advantage with respect to disease resistance (28-40% increase in genetic gain) and acted as an extra precaution against loss of the target QTL. However, for total merit index the advantage of gene-assisted genomic selection over genomic selection was lower (4-5% increase in genetic gain).

Conclusion: Substantial differences between introgression programs using classical and genomic selection were observed, and the former was generally inferior with respect to both genetic gain and the ability to preserve the target QTL. Combining genomic selection with gene-assisted selection for the target QTL acted as an extra precaution against loss of the target QTL and gave additional genetic gain for disease resistance. However, the effect on total merit index was limited.

Figure 1

Genetic levels for a) production trait (PT), b) disease resistance (DR) and c) total merit index (TMI) by generation for the different selection schemes under scenario 1; the selection schemes presented are; PRL, BACKCROSS CBLUP (BCL), BACKCROSS GBLUP (BGL) and BACKCROSS gene-assisted GBLUP (BGGL).

Figure 2

Allele frequencies of the favourable allele of the major QTL for the BACKCROSS selection scheme under a) classical (CBLUP), b) genomic (GBLUP), and c) gene-assisted genomic (GasGBLUP) selection in Scenario 1. Single dots represent individual replicates, red line average observed frequencies of the target QTL, purple line expected frequencies if the target QTL is not subject to selection and green line expected frequencies of "neutral" donor alleles.

Figure 3

Allele frequencies of the favourable allele of the major QTL for the BACKCROSS selection scheme under a) classical (CBLUP), b) genomic (GBLUP), and c) gene-assisted genomic (GasGBLUP) selection in Scenario 2. Single dots represent individual replicates, red line average observed frequencies of the target QTL, purple line expected frequencies if the target QTL is not subject to selection and green line expected frequencies of "neutral" donor alleles.