Utilisation d’une puce très basse densité (1 100 SNP) pour la sélection génomique chez 3 races porcines françaises

Céline  Carillier-Jacquin (Inrae) et al., 52e Journées de la Recherche Porcine (FRA), 4 et 5 février 2020, p. 45-46, poster

Poster.

Le coût actuel des puces SNP constitue une limite au développement de la sélection génomique dans les populations porcines (Badke et al., 2014; Wellmann et al., 2013). Afin de réduire les coûts de génotypage, une puce SNP basse densité (LD), conçue en 2016, est utilisée en routine. Ce panel LD d'environ 1100 SNP a été optimisé pour maximiser la précision d'imputation dans la population de porcs Landrace français. Dans la présente étude, nous avons proposé d’étudier l’impact de l’utilisation de ce panel pour deux autres grandes races de porc françaises, le Large White et le Piétrain. 

ENG

Poster.

Design of a very low density chip (1 100 SNP) for genomic selection in 3 French pig breeds 

To reduce genotyping costs for genomic selection, a Low-Density SNP (LD) chip, designed in 2016, is now used routinely. This panel is composed of approximately 1 100 equidistant SNPs. The relevance of this chip has been studied in French populations of the Landrace, Large White and Pietrain pig breeds. The quality of imputation was estimated by the correlation between actual and imputed genotypes and error rates. The impact of imputation on the genomic evaluations was estimated by the correlation between the genomic values obtained for the candidates with imputed genotypes, and those obtained with the high-density genotypes. Average error rates of imputation estimated on all the chromosomes were 0.03, 0.11 and 0.14 for Landrace, Large White and Pietrain, respectively. The estimated correlations between actual and imputed genotypes were relatively high at 0.93, 0.92 and 0.88 for Landrace, Large White and Pietrain populations, respectively. Correlations between genomic breeding values predicted with highdensity genomic data or imputed genomic data from the LD SNP panel ranged from 0.89-0.97 for Large White and Landrace populations for reproductive traits. They were higher than those obtained for the Pietrain population (0.80 and 0.97 for production traits, respectively). In conclusion, despite the limited number of SNPs on the low-density panel used in this study, the imputation accuracy is sufficient to use the imputed genotypes in the genomic evaluations. In practice, genotyping candidates with a LD chip is a solution for selecting future breeding pigs at lower cost.