Accuracy of genomic selection to improve litter traits in the French Landrace pig population

Poster.

• To assess gains in accuracy due to integration of genomic information in genomic evaluations of pigs,
• Focus on litter traits and on the French Landrace dam line.

abstract : 

The objective of this study was to assess the gain in selection accuracy obtained due to integration of genomic information in genetic evaluation models. In total, 579 boars (born>2002) and 504 sows (born>2010) were genotyped with the Porcine60K SNP panel. After edition, 1,067 animals were available for analysis. Four traits were analyzed: the number of piglets born alive (NBA), the number of weaned piglets (NW), the mean birth weight of piglets (MBW) and the within-litter standard deviation of piglet birth weights (SDBW). The predictive ability of the conventional BLUP and single-step GBLUP model was assessed with a validation study in which the genomic dataset was divided in a reference and a validation population. For the validation study, performances of individuals born after 12/31/2013 were removed. The reference population comprised 491 boars with daughters and 355 sows with own performances. The validation population was made up of 46 boars born in 2013 having at least 15 daughters with reproduction data on 09/30/2015. The predictive ability was estimated for validation boars as the correlation between the EBV estimated based on offspring performance and the EBV / GEBV estimated as candidates with performances recorded up to 12/31/2013.

We used the same fixed and random effects and genetic parameters as in the routine genetic evaluations. With conventional BLUP, the predictive ability of the model was 0.46 for NBA, 0.45 for NW, 0.37 for MBW and 0.26 for SDBW. With GBLUP model, the predictive ability was 0.55 for NBA, 0.63 for NW, 0.45 for MBW and 0.26 for SDBW. Thus, genomic evaluation improves the accuracy of selection for all maternal traits.