پردیس کشاورزی و منابع طبیعی دانشگاه تهران Iranian Journal of animal Science 2008-4773 43 2 2012 08 22 Comparison between Bayesc and GBLUP in Estimating Genomic Breeding Values under Different QTL Variance Distributions Comparison between Bayesc and GBLUP in Estimating Genomic Breeding Values under Different QTL Variance Distributions 261 268 28534 10.22059/ijas.2012.28534 FA Masoud Shirali Seyed Reza Miraei-Ashtiani Abbas Pakdel Chris Haley Ricardo Pong-Wong Journal Article 1970 01 01 A genome consisted of 1000 biallelic single nucleotide polymorphisems (SNPs) on one chromosome with 100 CM length was simulated and different QTL variance distributions (Uniform, Normal and Gamma) and various numbers of QTL (5, 10 and 20) were considered as simulation assumptions and consecutively 9 various traits were generated. The comparison between GEBVs obtained from BayesC and GBLUP showed that GEBVs and true breeding values were largely correlated (r > 0.80) in training population for all traits. Comparing accuracies of GEBVs showed that both BayesC and GBLUP methods performed similarly, except in traits with 5 QTLs BayesC performed significantly better (P < 0.05). The accuracies of estimations using BayesC indicated that this method performed better under scenarios with low number of QTL and gamma variance distribution in trait of interest. However, GBLUP presented accurate estimations in all traits; and number of QTL and QTL variance distributions had no impact on accuracies of GBLUP estimations. These results can be due to assumed genetics model in trait of interest in each method. A genome consisted of 1000 biallelic single nucleotide polymorphisems (SNPs) on one chromosome with 100 CM length was simulated and different QTL variance distributions (Uniform, Normal and Gamma) and various numbers of QTL (5, 10 and 20) were considered as simulation assumptions and consecutively 9 various traits were generated. The comparison between GEBVs obtained from BayesC and GBLUP showed that GEBVs and true breeding values were largely correlated (r > 0.80) in training population for all traits. Comparing accuracies of GEBVs showed that both BayesC and GBLUP methods performed similarly, except in traits with 5 QTLs BayesC performed significantly better (P < 0.05). The accuracies of estimations using BayesC indicated that this method performed better under scenarios with low number of QTL and gamma variance distribution in trait of interest. However, GBLUP presented accurate estimations in all traits; and number of QTL and QTL variance distributions had no impact on accuracies of GBLUP estimations. These results can be due to assumed genetics model in trait of interest in each method. https://ijas.ut.ac.ir/article_28534_bff5e7df581d52248ef25729b15b839b.pdf