Improving Genomic Evaluation of Holstein Cattle Using a Haplotype-Based Relationship Matrix

پژوهشهای علوم دامی ایران

Article 7, Volume 10, Issue 3 - Serial Number 35, September 2019, Pages 393-402   

Document Type: Genetics & breeding

DOI: 10.22067/ijasr.v10i3.66156

Authors

mehrnush forutan¹; Saeed Ansari Mahyari^{* 2}; Flavio Schenkel³; Mehdi Sargolzaei³

¹Isfahan University of Technology

²Faculty of Agriculture, Isfahan University of Technology, Isfahan, Iran

³University of Guelph

Abstract

Abstract Introduction With the advent of high throughput genotyping technologies, interest has grown in using genomic information to estimate breeding values. Genomic selection, in which genetic markers across the whole genome are used to estimate breeding values of individuals, is routinely applied in dairy cattle breeding programs. In dairy cattle, genomic selection has resulted in a substantial increase in the rate of genetic gain compared to traditional selection. This has been achieved mainly by reducing the generation interval, which became possible because of the higher accuracies of genomic breeding values (GEBV) estimated early in life compared to parent averages. Since single nucleotide polymorphism (SNP) genotypes are bi-allelic and, therefore, their information content is not high, so SNP-based methods may not effectively capture the linkage disequilibrium (LD) between SNPs and multi-allelic quantitative trait loci (QTLs). Haplotypes are in general “multi-allelic” and compared to individual SNPs may better capture LD with multi-allelic QTL. Furthermore, most of the SNPs in the chips are old mutations. This may imply that SNP-based relationship matrix traces very old relationships from distant relatives and, therefore, may not trace changes due to recent selection accurately. It has also been hypothesized that using similarity between haplotypes to model the covariance between genomic effects can result in better predictive ability than modeling covariance based on SNP genotypes. So the objective of this study was to investigate the accuracy and bias of GEBV using genomic best linear unbiased prediction (GBLUP) with alternate genomic relationship matrices (G) based on SNPs and haplotypes information. Materials and Methods The North American Holstein genotype data was provided by the Canadian Dairy Network (CDN). The Holstein bulls with official domestic proofs were genotyped using the Illumina Bovine SNP 50 TM Chip. The genotyped Holstein bulls were classified as estimation group and prediction group. All bulls in prediction group were those bulls who were born from 2007 to 2011 and had official proof in 2015. De-regressed EBV (dEBV) based on the 2015 genetic evaluation (dEBV2015) were used for validation purposes. The estimation group included bulls born mainly between 1960 to 2007. The analyzed traits were fat yield, milk yield, somatic cell score, conformation, and days open. The GEBVs for mentioned traits were estimated based on GBLUP models using SNP1101 software. Three G matrices were built: 1) SNP genotype based relationship matrix (GSNP), 2) haplotype based relationship matrix (GHAP), and 3) hybrid matrix composing of SNP genotype and haplotype based relationship matrices (GSNP, HAP). Accuracy of prediction was calculated as Pearson’s correlation between estimated GEBV and dEBV2015 for prediction group. Bias of prediction was also calculated as regression coefficient of dEBV2015 on GEBV for prediction group. Results and Discussion Observed differences between alternate G matrices were larger in prediction bias than in prediction accuracy. Accuracy of genomic predictions based on GHAP were 0.73, 0.71, 0.62, 0.54, and 0.50 for fat yield, milk yield, somatic cell score, conformation, and days open, respectively. Accuracy of genomic predictions based on GSNP also were 0.72, 0.70, 0.61, 0.56, and 0.49, respectively. Although using GHAP instead of GSNP did not significantly improve the accuracy of prediction, it resulted in estimates with 6 (conformation) to 15% (days open) less bias and closer to one over the different traits. Genomic selection based on GHAP instead of GSNP seems to improve bias of prediction for traits under recent selection, such as days open. This might also be due to the fact that this trait is lowly heritable. Although accuracy of genomic predictions based on GSNP, HAP were similar to GSNP, using GSNP, HAP resulted in higher biases than GHAP or GSNP. Conclusion Based on these findings, GHAP might be an alternative approach to reduce bias of genomic predictions in routine genomic evaluations. Genomic selection based on GHAP instead of GSNP improved bias of prediction especially for low heritable traits under recent selection, such as days open. Small gain achieved with GHAP compared to GSNP may imply that the correlation between IBS and IBD in GHAP is not optimal for maximization of prediction accuracy. Based on current results further research is needed to investigate the use of haplotype length and allele frequency of markers in haplotype segments in the definition of haplotype similarity. In addition, this study suggests further research to assess the effect of recent selection, heritability and genetic architecture of the traits on the performance of haplotype based relationship matrix.

Keywords

Accuracy; Bias of prediction; Genomic relationship matrix; Haplotype; Real data

References

Bohmanova, J., M. Sargolzaei, and F. S. Schenkel. 2010. Characteristics of linkage disequilibrium in North American Holsteins. BMC Genomics, 11:421.

Calus, M. P. L., T. H. E. Meuwissen, A. P. W. de Roos, and R. F. Veerkamp. 2008. Accuracy of genomic selection using different methods to define haplotypes. Genetics, 178: 553–561.

Dekkers, J. C. M. 2005. Animal Breeding Strategies. Available at http:// www.anslab.iastate.edu/Class/AnS652X/Chapter4.pdf.

de Roos, A. P. W., C. Schrooten, and T. Druet. 2011. Genomic breeding value estimation using genetic markers, inferred ancestral haplotypes, and the genomic relationship matrix. Journal of Dairy Science, 94:4708–4714.

Edriss, V., B. Guldbrandtsen, M.S. Lund, and G. Su. 2013. Effect of marker-data editing on the accuracy of genomic prediction. Journal of Animal Breeding and Genetics, 130: 128–135.

Ferdosi, M. H., J. Henshall, and B. Tier. 2016. Study of the optimum haplotype length

to build genomic relationship matrices. Genetic Selection Evolution, 48:75.

Habier, D., R. L. Fernando, and J. C. M. Dekkers. 2007. The impact of genetic relationship information on genome-assisted breeding values. Genetics, 177: 2389–2397.

Habier, D., R. L. Fernando, and D. J. Garrick. 2013. Genomic BLUP decoded: a look into the black box of genomic prediction. Genetics, 194:597–607.

Hickey, J. M., B. P. Kinghorn, B. Tier, S. A. Clark, J. H. J. van der Werf, and G. Gorjanc. 2013. Genomic evaluations using similarity between haplotypes. Journal of Animal Breeding and Genetic, 130: 259-269.

Meuwissen, T. H. E., B. J. Hayes, and M. E. Goddard. 2001. Prediction of total genetic value using genome-wide dense marker maps. Genetics, 157:1819–1829.

Meuwissen, T. H., J. Odegard, I. Andersen-Ranberg, and E. Grindflek. 2014. On the distance of genetic relationships and the accuracy of genomic prediction in pig breeding. Genetic Selection Evolution, 46:49.

Miglior, F., B. L. Muir, and B. J. Van Doormaal. 2005. Selection Indices in Holstein Cattle of Various Countries. Journal of Dairy Science, 88:1255–1263.

Nejati Javaremi, A., C. Smith, and J. P. Gibson. 1997. Effect of total allelic relationship on accuracy of evaluation and response to selection. Journal of Animal Science, 7: 1738-1745.

Parker Gaddis, K. L., J. B. Cole, J. S. Clay, and C. Maltecca. 2014. Genomic selection for producer-recorded health event data in US dairy cattle. Journal of Dairy Science, 97:3197–3199.

Sargolzaei, M., and F. S. Schenkel. 2009. QMSim: A large-scale genome simulator for livestock. Bioinformatics, 25(5):680-681.

Sargolzaei, M., 2014. SNP1101 User’s Guide. Version 1.0. HiggsGene Solutions Inc., Guelph, Ontario, Canada.

Schaeffer, L. R. 2006. Strategy for applying genome wide selection in dairy cattle. Journal of Animal Breeding and Genetic, 123(4):218-223.

Su, G., O. F. Christensen, L. Janss, and M. S. Lund, 2014. Comparison of genomic predictions using genomic relationship matrices built with different weighting factors to account for locus-specific variances. Journal of Dairy Science, 97 :6547–6559.

VanRaden, P. M. 2008. Efficient methods to compute genomic predictions. Journal of Dairy Science, 91:4414–4423.

VanRaden, P. M., C. P. Van Tassell, G. R. Wiggans, T. S. Sonstegard, R. D. Schnabel, J. F. Taylor, and F. S. Schenkel. 2009. Invited review: Reliability of genomic predictions for North American Holstein bulls. Journal of Dairy Science, 92:16–24.