نوع مقاله : مقاله پژوهشی
نویسنده
گروه علوم دامی- ژنتیک و اصلاح نژاد دام وطیور- دانشکده کشاورزی- دانشگاه لرستان- خرم آباد- ایران
چکیده
کلیدواژهها
موضوعات
عنوان مقاله [English]
نویسنده [English]
BayesianLASSO is a penalized regression method with the potential of variable selection using the regularization parameter λ. In a fully Bayesian framework, it is possibleto treat λ as random parameter and different priors can be assigned to it. The effect of assigning three different priors including gamma, beta, and fixed distributions on the predictability and estimation of genetic parameters was investigated using mice genome and seven traits with different genetic architecture. The results showed that estimation of variance components and heritabilityare strongly affected by the assigned prior.Highest genetic variance and heritabilityand the lowest residual variance were estimated by fixed λ model for different traits. The average accuracy obtained for each trait using different priors was relatively similar, but standard deviation of the accuracies by the fixed λ model was always greater than the other two priors. Comparison of the minimum and maximum values of the models accuracies via fivefold cross validation revealed that the highest accuracy values were obtained using the assignment of gamma and beta distributions to the regularization parameter. By increasing the heritabilitynd the number of background QTLof the traits, the accuracy of the models increased and the difference between the three different priors decreased.Thedifference between the models was greater regarding to prediction bias, but it generally followed the same trend of the accuracy, i.e. the models with the highest accuracy also had the lowest prediction bias. Based on the presentresults, proper prior assignment to λ is more necessary for traits under small number of QTL and lowheritability.
کلیدواژهها [English]
)