تعداد نشریات | 295 |

تعداد شمارهها | 3,652 |

تعداد مقالات | 28,308 |

تعداد مشاهده مقاله | 20,685,102 |

تعداد دریافت فایل اصل مقاله | 12,569,596 |

## The Impact of Different Genetic Architectures on Accuracy of Genomic Selection Using Three Bayesian Methods | ||

Iranian Journal of Applied Animal Science | ||

مقاله 7، دوره 8، شماره 1، خرداد 2018، صفحه 53-59
اصل مقاله (522 K)
| ||

نوع مقاله: Research Articles | ||

نویسندگان | ||

F. Alanoshahr ^{} ^{1}؛ S.A. Rafat^{1}؛ R. Imany-Nabiyyi^{2}؛ S. Alijani^{1}؛ C. Robert Granie^{3}
| ||

^{1}Department of Animal Science, Faculty of Agriculture, University of Tabriz, Tabriz, Iran | ||

^{2}Department of Animal Science, Faculty of Mathematical Science, University of Tabriz, Tabriz, Iran | ||

^{3}INRA-INPT-ENSAT-INPT-ENVT, Université de Toulouse, UMR 1388 GenPhySE, Castanet Tolosan, France | ||

چکیده | ||

Genome-wide evaluation uses the associations of a large number of single nucleotide polymorphism (SNP) markers across the whole genome and then combines the statistical methods with genomic data to predict the genetic values. Genomic predictions relieson linkage disequilibrium (LD) between genetic markers and quantitative trait loci (QTL) in a population. Methods that use all markers simultaneously may therefore result in greater reliabilities of predictions of the total genetic merit, indicating that a larger proportion of the genetic variance is explained. This is hypothesized that the genome-wide methods deal differently with genetic architecture of quantitative traits and genome. The genomic nonlinear Bayesian variable selection methods (BayesA, BayesBand Bayesian LASSO) are compared using the stochastic simulation across three effective population sizes (Ne). Thereby, a genome with three chromosomes, 100 cM each was simulated. For each animal, a trait was simulated with heritability of 0.50, three different marker densities (1000, 2000 and 3000 markers) and number of the QTL was assumed to be either 100, 200 or 300. The data were simulated with two different QTL distributionswhich were uniform and gamma (α=1.66, β=0.4). Marker density, number of the QTL and the QTL effect distributions affected the genomic estimated breeding value accuracy with different Ne (P<0.05). In comparison of three methods, the greatest genomic accuracy obtained by BayesB method for traits influenced by a low number of the QTL, high marker density, gamma QTL distribution and high Ne. | ||

کلیدواژهها | ||

BayesA؛ BayesB؛ genomic accuracy؛ LASSO؛ marker density؛ Ne | ||

اصل مقاله | ||

The genome-wide evaluation combines traditional approaches for prediction of genetic values with using high throughput genotype data such as SNP (Meuwissen
The populations were simulated using the QMSim softwere (Sargolzaei and Schenkel, 2009) based on forward-in-time process. A genome consisted of three chromosomes with a length of 100cM was simulated 1000, 2000 and 3000 SNPs were equally spaced over the chromosomes. Three different numbers of QTL (100, 200 and 300) were considered and QTLs were uniformly distributed over the chromosomes. One hundred individuals, including 50 males and 50 females, were simulated for the base population (zero generation). These loci were assumed to be biallelic for both SNPs and QTL with allele frequencies equal to 0.50 (Table 1). The simulation started with an initial population of 100 N Where: a m: total number of QTL. z
The LD measure r Where: D= f(AB) − f(A).f(B). f(AB), f(A), f(a), f(B), f(b): frequencies of haplotypes AB and of alleles A, a, B, b, respectively (Meuwissen
Three methods, BayesA, BayesB and Bayes LASSO, were used to estimate QTL, SNPs effects and genomic breeding values. The main difference between these three applied approaches is in their assumptions regarding genetic models of the trait. The genomicestimated breeding values (GEBV) for individuals in validation generations for threeBayesA, BayesB and Bayes LASSO methods were predicted using the model (Meuwissen Where: n: number of SNPs across the genome. X g
In this model like GBLUP, all SNPs are assumed to have some effects, however, assumed that some of the SNPs are in LD with QTL of moderate to large effects. The SNP effects sampled from a normal distribution with the variance for each SNP sampled from an inverse scaled Chi-square distribution. The BayesA method was performed using the model (Meuwissen Where: S: scale parameter. v: number of degrees of freedom.
In BayesB assumes that many of the SNPs are in genomic regions where there are no QTL and thus have zero effects, whilst a small proportion of SNPs are in LD with QTL and consequently do have an effect. This structure means that those effects that are non-zero can be thought of as those in stronger LD with the QTL. In fact, if the number of times a SNP is included in the model (
The Bayes LASSO, where a large proportion of marker effects are set to zero and the Bayesian LASSO, where marker effects are modeled using a double exponential distribution, with a high peak at zero and heavy tail that accommodate SNPs with larger effects. Each marker has the same double exponential distribution and no heterogeneous variance either (Park and Casella, 2008). Better estimates are obtained where many possible QTL are estimated to have zero effect or, equivalently, excluded from the model. If all the QTL effects were from a reflected exponential distribution (i.e. without extra weight at zero), an estimator called the LASSO is the appropriate one (Tibshirani, 1996). However, in the situation where many true effects are zero, LASSO still estimates too many nonzero effects. A pragmatic alternative is to exclude from the model but the most highly significant effects. The λ parameter in the LASSO approach was assigned a gamma(a, b) prior distribution. Values of a and b were set at 0.05 and 1.0, respectively, so that prior of λ was essentially uniform over a wide range of values. Steps of the algorithm are outlined below (Yi and Xu, 2008): For each analysis, a markov chain monte carlo (MCMC) with 210000 cycles with WinBUGS1.4 software ran and the first 10000 cycles were discarded as burn-in period. Estimates at every 5
The effects of heritability levels, marker density panels, and the number of QTLs on the accuracy of genomic predictions were evaluated using PROC GLM, and the average accuracies of GEBV were compared using the least squares means (LSM) procedure at P < 0.05 (SAS, 2003). The correlation between the GEBV and true genomic breeding value (TGBV) was used as measure of accuracy.
This study demonstrates the accuracy of genomic selection with different numbers of the QTL, marker densities and QTL effect distribution in different N
The relative genomic accuracy increased with increasing of marker densities and N
The reason can be attributed to increase the number of known data (number of phenotypic records in the base population) versus the number of unknowns variables (SNP effects). When the number of observations in the base population are greater, the SNP effects more precisely estimate and eventually genomic breeding values will be greater accurate (Daetwyler
Increasing the numbers of the QTL from 100 to 300, decreased the average genomic accuracy in all three N
The BayesB method was accurate in comparison with both BayesA (P<0.05) and Bayesian LASSO (P>0.05) methods. Among the three methods, the greatest genomic accuracy obtained in low numbers of the QTL (100), high marker density, gamma QTL effect distribution, and large numbers of N
The results of this study are in agreement with Daetwyler
The extent of LD have major impact on the accuracy of the MEBV. Based on the findings of this simulation study, low QTL number, as well as high dense marker panels, aiming to increase the level of LD between markers and the QTL, will likely be needed for successful implementation of the genomic selection. To implement genomic selection with the LD panels, a training population of sufficient size is necessary. Using a dense marker map covering all chromosomes, it is possible to accurately estimate the breeding value of animals that have no phenotypic record of their own. The BayesB produced estimates with greater accuracies in traits influenced by low number of the QTL and with the gamma QTL effect distribution. Also in the low N
The authors thank C. Robert Granie for encouraging us to write this article and for comments provided on earlier versions of this manuscript. | ||

مراجع | ||

Daetwyler H.D., Pong-Wong R., Villanueva B. and Woolliams J.A. (2010). The impact of genetic architecture on genome-wide evaluation methods. Genetics. 185, 1021-1031.
Daetwyler H.D., Villanueva B., Bijma P. and Woolliams J.A. (2007). Inbreeding in genome-wide selection. J. Anim. Breed. Genet.124, 369-376.
De los Campos G., Naya H ., Gianola D., Crossa J., Legarra A., Manfredi E., Weigel K. and Cotes J.M. (2009). Predicting quantitative traits with regression models for dense molecular markers and pedigree. Genetics. 182, 375-385.
Gianola D. and van Kaam J. (2008). Reproducing kernel Hilbert spaces regression methods for genomic assisted prediction of quantitative traits. Genetics. 178(4), 2289-2303.
Goddard M. (2009). Genomic selection: Prediction of accuracy and maximisation of long term response. Genetics. 136, 245-257.
Habier D., Fernando R.L., Kizilkaya K. and Garrick D.J. (2011). Extension of the Bayesian alphabet for genomic selection. BMC Bioinform. 12, 186-193.
Haldane J.B.S. (1919). The combination of linkage values and the calculation of distances between the loci of linked factors. Genetics. 8, 299-309.
Hill W.G. and Robertson A. (1968). Linkage disequilibrium in finite populations. Theor. Appl. Genet. 38, 226-231.
Meuwissen T.H.E., Hayes B.J. and Goddard M.E. (2001). Prediction of total genetic value using genome-wide dense marker maps. Genetics. 157, 321-322.
Nadaf J. and Pong-Wong R. (2011). Applying different genomic evaluation approaches on QTLMAS2010 dataset. BMC Proc. 5(3), 9-16.
Park T. and Casella G. (2008). The Bayesian LASSO. J. Am. Stat. Assoc. 103, 681-686.
Sargolzaei M. and Schenkel F.S. (2009). QMSim: A large-scale genome simulator for livestock. Bioinformatics. 25, 680-681.
SAS Institute. (2003). SAS ^{®}/STAT Software, Release 9.1. SAS Institute, Inc., Cary, NC. USA.
Shirali M., Miraei-Ashtiani S.R., Pakdel A., Haley C. and Pong-Wong R. (2015). A comparison of the sensitivity of the BayesC and genomic best linear unbiased prediction (GBLUP) methods of estimating genomic breeding values under different quantitative trait locus (QTL) model assumptions. Iranian J. Appl. Anim. Sci. 5(1), 41-46.
Solberg T.R., Sonesson A.K., Woolliams J.A. and Meuwissen T.H.E. (2008). Genomic selection using different marker types and densities. J. Anim. Sci. 86, 2447-2454.
Sved J.A. (1971). Linkage disequilibrium and homozygosity of chromosome segments in finite populations. Theor. Popul. Biol. 2, 125-141.
Tibshirani R. (1996). Regression shrinkage and selection via the LASSO. J. Roy. Stat. Soc. B Met. 58, 267-288.
Whittaker J.C., Thompson R. and Denham M.C. (2000). Marker-assisted selection using ridge regression. Genet. Res. 75(2), 249-252.
Wimmer V., Lehermeier C., Albrecht T., Auinger H.J., Wang Y. and Schön C.C. (2013). Genome-wide prediction of traits with different genetic architecture through efficient variable selection. Genetics. 195, 573-587.
Yi N. and Xu S. (2008). Bayesian LASSO for quantitative trait loci mapping. Genetics. 179, 1045-1055. | ||

آمار تعداد مشاهده مقاله: 52 تعداد دریافت فایل اصل مقاله: 19 |
||