Comparison of accuracy of genome-wide and BLUP breeding value estimates in sib based aquaculture breeding schemes

Summary

This study compared the accuracy for genome-wide breeding values (GWEBV) with traditional BLUP breeding values estimated from phenotypic performance of sibs (BLUPEBV) in a typical sib based aquaculture breeding scheme. The breeding schemes were closed nuclei with 1000 selection candidates from 100 families in each generation G1 to G5. Phenotypes were recorded on sibs of the selection candidates. Both selection candidates and their recorded sibs were genotyped using dense SNP genotyping technology and marker effects estimated from the sibs were used to predict GWEBV of the selection candidates. With 40 sibs, accuracy in G3 was 0.54 for BLUPEBV compared to 0.78 for GWEBV (marker density 0.5N(e)/M), i.e. a relative increase of 44%. When the number of sibs for the GWEBV scheme decreased from 40 to 20 sibs, accuracy decreased by 7% in G3. Both for BLUPEBV and GWEBV breeding schemes, the effect of heritability on accuracy was small and accuracy increased around 4% when increasing heritability from 0.2 to 0.4. Contrary to expectations differences in accuracy between BLUPEBV and GWEBV were the same for the low (0.2) and high (0.4) heritability, which may be due to the large number of phenotypic records used. When records were only available in generation G1, accuracy of GWEBV decreased over generations and was 0.66 in G1, 0.57 in G2 and 0.51 in G3 with 40 sibs and marker density 0.5N(e)/M. Accuracy increased with increased marker density for GWEBV, but the increase was small when marker density increased from 0.5 to 1.0N(e)/M except when the test group size was increased from 20 to 40, indicating that with a higher marker density a higher number of phenotypes is needed. Increasing the size of the genome (1000 cM vs. 2000 cM) decreased the accuracy for the GWEBV breeding values but increased the accuracy for the BLUPEBV breeding values. Overall, accuracy was up to 33% higher for GWEBV than for BLUPEBV for the studied scenarios because both between and within family genetic variances are utilized when estimating genome-wide breeding values. The results indicate that aquaculture breeding companies can increase the accuracy of selection when using genomic selection in a typical sib based aquaculture breeding scheme. (C) 2009 Elsevier B.V. All rights reserved.