نوع مقاله : مقاله پژوهشی
نویسندگان
1 دانشیار، گروه علوم دامی، دانشکده کشاورزی و منابع طبیعی، دانشگاه اراک
2 استادیار، گروه علوم دامی، دانشکده کشاورزی و منابع طبیعی، دانشگاه اراک
چکیده
کلیدواژهها
موضوعات
عنوان مقاله [English]
نویسندگان [English]
Introduction: The goal of genome-wide association (GWA) studies of quantitative traits is to identify genomic regions that explain a substantial proportion of the genetic variation for the trait, with the ultimate goal to identify causal mutations underlying the genetic basis of the trait. The standard GWA approach is to genotype a population that has been phenotyped for the trait(s) of interest and genotyped for many genetic markers across the genome and to analyze these data by estimating and testing the effects of marker genotypes on phenotypes using a regression-type of analysis for each single nucleotide polymorphism (SNP), one at a time. Bayesian methods such as Bayes A and Bayes B assume a heavy tail prior distribution for SNP effects and use Markov Chain Monte Carlo (MCMC) to sample from the posterior distribution. Although the objective of these methods was to predict the breeding value of selection candidates (genomic breeding values), they do that by estimating the effects of all SNPs. The estimated SNP effect, the proportion of variance explained by a SNP, or the number of times the SNP fits in the model with non–zero effect can be used as criteria to identify locations or genomic regions that affect the trait of interest. Results have shown that these Bayesian methods can effectively detect QTL in simulated and real data. Recently, a new methodology has been developed to address this limitation and allow for a better understanding of the genetic architecture of complex traits through a gene network analysis. For this purpose, to identify genomic regions and candidate genes associated with egg weight (EW), a genome-wide association study (GWAS) was performed in the present study using Affymetrix 600 K high density SNP array in 1,078 hens of 11th generation of Rhode Island Red.
Materials and methods: Data available for 1,078 pedigree-recorded hens were used to collect phenotypic EW-related data. Seven traits, including egg weight at the first laying of hens, and egg weight at 28, 36, 56, 66, 72, and 80 weeks of age were collected for each bird. The analyses were performed using GenSel v4.73R, by fitting covariates for haplotype alleles in BayesA and BayesB models. A single Markov chain Monte Carlo (MCMC) chain of length 41,000, including burn-in of 1,000 first iterations, was computed for each analysis to obtain posterior estimates of covariate effects. These were used to obtain a direct genetic variance for animals. The primary analysis showed that correlations and regression coefficients had converged at this chain length. Annotation terms and pathway analyses were conducted using protein analysis through evolutionary relationships of PANTHER software version 10.0.
Results and discussion: The results showed that the BayesA method performed better in explaining additive genetic variance compared to BayesB method. Nine markers obtained from BayesA with the highest additive genetic variance were located on chromosomes 1, 3, 5, and 20. Genes that overlap in regions of interest were identified with the Ensembl BioMart data mining (http://www.ensembl.org/biomart/) based on the Galgal6 assembly and the Ensembl Genes 96 database. The detected SNPs were located close to 35 genes, among which, the candidate genes of BPIFB2, OCX36, CPT1A, TCF15, CECR2, SIAH3, FADS1, FADS2, and SGK1 play important functions in the egg production process through the albumen protein formation, fatty acids metabolism, and eggshell formation. It is noteworthy that the present study has detected an association in regions different from that reported by previous studies. This can be because of flock particularities, such as the extent of linkage disequilibrium, allelic frequencies, and statistical approaches.
Conclusions: The results of the present study showed that when the genetic architecture of studied traits follows infinitesimal model assumptions, the BayesA method usually performs better than BayesB. Moreover, considering the identification of new genome regions and the key role of the mentioned genes on the development of egg weight, the efficiency of the BayesA method can be confirmed for GWAS in egg weight traits.
کلیدواژهها [English]
)