Evaluation of Bayesian alphabet and GBLUP based on different marker density for genomic prediction in Alpine Merino sheep

Abstract

The marker density, the heritability level of trait and the statistical models adopted are critical to the accuracy of genomic prediction (GP) or selection (GS). If the potential of GP is to be fully utilized to optimize the effect of breeding and selection, in addition to incorporating the above factors into simulated data for analysis, it is essential to incorporate these factors into real data for understanding their impact on GP accuracy, more clearly and intuitively. Herein, we studied the genomic prediction of six wool traits of sheep by two different models, including Bayesian Alphabet (BayesA, BayesB, BayesC π and Bayesian LASSO) and genomic best linear unbiased prediction (GBLUP). We adopted 5-fold cross-validation to perform the accuracy evaluation based on the genotyping data of Alpine Merino sheep (n = 821). The main aim was to study the influence and interaction of different models and marker densities on GP accuracy. The GP accuracy of the six traits was found to be between 0.28 and 0.60, as demonstrated by the cross-validation results. We showed that the accuracy of GP could be improved by increasing the marker density, which is closely related to the model adopted and the heritability level of the trait. Moreover, based on two different marker densities, it was derived that the prediction effect of GBLUP model for traits with low heritability was better; while with the increase of heritability level, the advantage of Bayesian Alphabet would be more obvious, therefore, different models of GP are appropriate in different traits. These findings indicated the significance of applying appropriate models for GP which would assist in further exploring the optimization of GP.

Improvement of Genomic Predictions in Small Breeds by Construction of Genomic Relationship Matrix Through Variable Selection

“…In our study, using both simulated and real data sets, we demonstrated that the accuracy of (ss)GBLUP increases when it is performed when the number of SNPs to construct G was reduced. This finding agrees with that of the extensive literature supporting the increased accuracy of Bayesian variable selection models in many different species (Lourenco et al, 2014; Mehrban et al, 2021 ; Yoshida et al, 2018; Zhu et al, 2021 ). For example, Akbarzadeh et al (2021) integrated only a subset of chosen SNPs into the GBLUP framework based on a classical GWAS analysis (i.e., 1, 5, 10, and 50% of significant SNPs).…”

“…As mentioned above, Bayesian SNP regression, or (ss)GBLUP using a weighted realized relationship matrix (Tiezzi and Maltecca, 2015; Zhang et al, 2016 ), always provides higher prediction accuracy than models assuming homogenous variance among SNPs (GBLUP or SNP-BLUP). However this increase in accuracy is often connected with increases in bias, especially when time cross-validation is used ( Mehrban et al, 2021 ), instead of five-fold or leave-one-out cross-validation ( Zhu et al, 2021 ). However, when the goal is to achieve the “best predictor”, namely, a value closer as possible to real one, models assuming heterogenous variances and models with variable selection can be identified as the best models.…”

“…Moreover, it is unclear how models considering heterogenous variances account for selection since only k-fold cross-validation is usually applied ( Zhu et al, 2021 ). In real-life breeding scenarios, time cross-validation should be considered ( Liu, 2010 ) because this validation method mimics the true accumulation of information across time.…”

Using Pedigree and Genomic Data toward Better Management of Inbreeding in Italian Dairy Sheep and Goat Breeds

“…As previously mentioned, genomic inbreeding coefficients are more reliable than pedigree-based ones, being free from possible registration inaccuracy [ 4 , 11 ]. This is the reason why the implementation of genomic tools at the field level has increased, leading to genomic selection programs not only in cattle but also in other farm animals such as sheep and goats [ 31 , 32 , 33 ].…”

403. Predicting pig digestibility coefficients with microbial and genomic data using machine learning prediction algorithms

“…Genetic evaluation as a sum of additive trait effects using genomic Best Linear Unbiased Predictor (BLUP) method, as defined by Meuwissen et al (2001), is known to be more accurate than other methods for polygenic traits in animal genetics (Zhu et al, 2021;Wang et al, 2019). Nevertheless, some non-parametric machine learning methods allow considering interacting genes, or interaction between microbiota and host genetic and major effect of some gene or OTUs (Operational Taxonomic Unit).…”