Study of QTL Effects Distribution on Accuracy of Genomic Breeding values Estimated Using Bayesian Method

Document Type : Genetics & breeding

Authors

• nazanin mahmoudi ¹
• Ahmad Ayatollahi Mehrgardi ¹
• Mahmood Honarvar ²
• Ali Esmaili zadeh kashkuei ³

¹ Shahid Bahonar University of Kerman

² Islamic Azad University Share Qods

³ Department of Animal Sciences, Faculty of Agriculture, Shahid Bahonar University of Kerman, Kerman, Iran

10.22067/ijasr.v7i3.36312

Abstract

Introduction Genetic evaluation and estimation of breeding value are one of the most fundamental elements of breeding programmes for genetic improvement. Recently, genomic selection has become an efficient method to approach this aim. The accuracy of estimated Genomic breeding value is the most important factor in genomic selection. Different studies have been performed addressing the factors affecting the accuracy of estimated Genomic breeding value. The aim of this study was to evaluate the effect of beta and gamma distributions on the accuracy of genetic evaluation.
Materials and Methods A genome consisted of 10 chromosomes with 200 cm length was simulated. Markers were spaced on 0.2 cm intervals and different numbers of QTL with random distribution were simulated. Only additive gene effects were considered. The base population was simulated with an effective size of 100 animals and this structure continued up to generation 50 to creating linkage disequilibrium between the markers and QTL. The population size was increased to 1000 animals in generation 51 (reference generation). Marker effects were calculated from the genomic and phenotypic information. Genomic breeding value was computed in generations 52 to 57 (training generation). Effects of gamma 1 distribution (shape=0.4, scale=1.66), gamma 2 distribution (shape=0.4, scale=1) and beta distribution (shape1=3.11, shape2=1.16) were studied in the reference and training groups. The heritability values were 0.2 and 0.05.
Results and Discussion The results showed that accuracy of genomic breeding value reduced with passing generation (from 51 to 57) for two gamma distributions and beta distribution; this decrease may be due to two factors: recombination has negative impact on accuracy of genomic breeding value and selection reduces genetic variance as the number of generations increases. Accuracy of genomic estimated breeding value increased as the heritability increased so that the high heritability had more accuracy than low heritability in same QTL number. Number and distribution of genes is an important factor for accuracy of estimated breeding value. Duncan test was conducted by SPSS software. Results illustrated that there was no significant difference between the different distributions. Comparing accuracy of estimated breeding value showed that in the low heritability scenario with 10 and 20 QTL, gamma distribution 2 and gamma distribution 1 performed well, respectively, whilst in 50 and 100 QTL scenario, beta distribution was superior in both Lasso and Ridge methods. In the high heritability scenario with 50, 100 QTL gamma distributions 2 were superior in both Lasso and Ridge methods. With four QTL (10, 20, 50 & 100), in high heritability scenario, estimated genomic breeding value was often increased by increasing the number of QTL. This may be due to increasing linkage disequilibrium between markers and QTL. In general, the gamma distribution led to the increased accuracy of the estimations in both Lasso and Ridge methods.
Conclusion Marker density, method to estimate marker effects, QTL distribution, number of QTL, number of generations and trait heritability are some effective factors on accuracy of estimated genomic breeding value. The accuracy of estimated genomic breeding value is output of these factors and the distribution of genes is an important factor for accuracy of estimated genomic breeding value. We can conclude that, accuracy is reduced with increasing number of generations from base population to training population while the accuracy of estimated genomic breeding value is increased when breeding value of the reference group is used in lieu of the phenotypic records. In addition, accuracy of estimated genomic breeding value is enhanced by increasing heritability, so that, between three the distributions simulated in high heritability scenario, gamma 2 distribution increased accuracy of the estimates. Although, the size and distribution of QTL effects still greatly influence the effectiveness of the genomic prediction methods, but as a suggestion, models of genetic variation of genomic assessment should be considered, since a method of estimating breeding value may have (or produce) a better estimation with a specific model.

Keywords

• Accuracy
• breeding value
• Genomic Selection
• QTL distribution

---

References

1. Calus, M. P. L., T. H. E. Meuwissen, A. P. W. de Roos, and R. F.Veerkamp. 2007. Accuracy of genomic selection using different methods to define haplotypes. Genetics, 178:553-561.

2. Foroutani Far. S., H. Mehrabani Yeganeh, and M. Moradi Sharbabak, 2012. Comparision of the Accuracy of the Estimated Traditional and Genomic Breeding Values using Single and Multi-Trait Analyses. Iranian Journal of Animal Science, 43(4): 497-504. (In Persian)

3. Goddard, M. 2008. Genomic selection: prediction of accuracy and maximisation of long term response. Genetica, 136:245-257.

4. Hayes, B., M. E. Goddard. 2001. The distribution of the effects of genes affecting quantitativetraits in livestock.Genetics Selection Evolution, 33: 209–229.

5. Hayes, B., P. Bowman, A. Chamberlain, and M. Goddard. 2009. Invited review: Genomic selection in dairy cattle: Progress and challenges. Journal of Dairy Science, 92: 433-443.

6. Karimi. D., M. Tahmoorespoor, M. Dadpasand, A. Aslaminejad, and M. Sando Lund. 2014. The effect of increasing female number of reference population and imputed markers on reliability of genomic prediction. Iranian Journal of Animal Science Research, 6 (4): 270-278. (In Persian)

7. Kolbehdari, D., L. R. Shaeffer, and J. A. B. Robinson. 2007. Estimation of genome wide haplotype effect in half sib designs. Journal of Animal Breeding and Genetics, 124:356-361.

8. Long, N., D. Gianola, G. J. M. Rosa, K. A. Weigel, and S. Avendaٌo. 2007. Machine learning classification procedure for selecting SNPs in genomic selection: application to early mortality in broilers. Journal of Animal Breeding and Genetics, 124: 377 – 389.

9. 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.

10. Mrode, R. A. 2005. Linear models for the prediction of animal breeding values, 2ndedition. CABI, UK.

11. Muir, W.M. 2007. Comparison of genomic and traditional BLUP-estimated breeding value accuracy and selection response under alternative trait and genomic parameters. Journal of Animal Breeding and Genetics, 124: 342-355.

12. 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, 75: 1738-1745.

13. Shirali. M., R. \Miraie Ashtiani, A. Pakdel, C. Healy, and R. Pong-Wong. 2012. Comparison between Bayesc and GBLUP in Estimating Genomic Breeding Values under Different QTL Variance Distributions. Iranian Journal of Animal Science, 43(2): 261-268. (In Persian)

14. Tibshirani, R. 1996. Regression shrinkage and selection via the Lasso. J. the Royal Statistical Society Series B: Methodological, 58: 267.288.

15. Villanueva, B., R. Pong-Wong, J. Fernandez, M. A. Toro. 2005. Benefits from marker-assisted selection under an additive polygenic genetic model. Journal of Animal Science, 83: 1747-1752.

16. Zargarian. B., M. Amin Afshar, M. Saatchi, and A. Noushtari. 2010. Effect of increasing the density of the markers on the accuracy of predicted breeding values genomic. Journal of Animal and Environment, 2(1), 37-44. (In Persian).

17. Zhang, W., C. Smith. 1993. Simulation of marker-assisted selection utilizing linkagedisequilibrium: the effects of several additional factors. Theoretical and Applied Genetics, 492–496.