Genomic prediction of crossbred performance based on purebred Landrace and Yorkshire data using a dominance model - Archive ouverte HAL Accéder directement au contenu
Article Dans Une Revue Genetics Selection Evolution Année : 2016

Genomic prediction of crossbred performance based on purebred Landrace and Yorkshire data using a dominance model

Hadi Esfandyari
  • Fonction : Auteur correspondant
  • PersonId : 984631

Connectez-vous pour contacter l'auteur
Piter Bijma
Mark Henryon
  • Fonction : Auteur
  • PersonId : 980289
Ole Fredslund Christensen
  • Fonction : Auteur
Anders Christian Sørensen
  • Fonction : Auteur
  • PersonId : 980292

Résumé

Background In pig breeding, selection is usually carried out in purebred populations, although the final goal is to improve crossbred performance. Genomic selection can be used to select purebred parental lines for crossbred performance. Dominance is the likely genetic basis of heterosis and explicitly including dominance in the genomic selection model may be an advantage when selecting purebreds for crossbred performance. Our objectives were two-fold: (1) to compare the predictive ability of genomic prediction models with additive or additive plus dominance effects, when the validation criterion is crossbred performance; and (2) to compare the use of two pure line reference populations to a single combined reference population.Methods We used data on litter size in the first parity from two pure pig lines (Landrace and Yorkshire) and their reciprocal crosses. Training was performed (1) separately on pure Landrace (2085) and Yorkshire (2145) sows and (2) the two combined pure lines (4230), which were genotyped for 38 k single nucleotide polymorphisms (SNPs). Prediction accuracy was measured as the correlation between genomic estimated breeding values (GEBV) of pure line boars and mean corrected crossbred-progeny performance, divided by the average accuracy of mean-progeny performance. We evaluated a model with additive effects only (MA) and a model with both additive and dominance effects (MAD). Two types of GEBV were computed: GEBV for purebred performance (GEBV) based on either the MA or MAD models, and GEBV for crossbred performance (GEBV-C) based on the MAD. GEBV-C were calculated based on SNP allele frequencies of genotyped animals in the opposite line.Results Compared to MA, MAD improved prediction accuracy for both lines. For MAD, GEBV-C improved prediction accuracy compared to GEBV. For Landrace (Yorkshire) boars, prediction accuracies were equal to 0.11 (0.32) for GEBV based on MA, and 0.13 (0.34) and 0.14 (0.36) for GEBV and GEBV-C based on MAD, respectively. Combining animals from both lines into a single reference population yielded higher accuracies than training on each pure line separately. In conclusion, the use of a dominance model increased the accuracy of genomic predictions of crossbred performance based on purebred data.
Fichier principal
Vignette du fichier
12711_2016_Article_220.pdf (1 Mo) Télécharger le fichier
12711_2016_220_MOESM1_ESM.docx (18.74 Ko) Télécharger le fichier
12711_2016_220_MOESM2_ESM.docx (30.01 Ko) Télécharger le fichier
12711_2016_220_MOESM3_ESM.docx (20.36 Ko) Télécharger le fichier
Origine : Publication financée par une institution
Origine : Publication financée par une institution
Origine : Publication financée par une institution
Origine : Publication financée par une institution
Loading...

Dates et versions

hal-01341374 , version 1 (04-07-2016)

Identifiants

Citer

Hadi Esfandyari, Piter Bijma, Mark Henryon, Ole Fredslund Christensen, Anders Christian Sørensen. Genomic prediction of crossbred performance based on purebred Landrace and Yorkshire data using a dominance model. Genetics Selection Evolution, 2016, 48 (1), pp.40. ⟨10.1186/s12711-016-0220-2⟩. ⟨hal-01341374⟩
23 Consultations
76 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More