The aim of this study was to compare the predictive performance of ridge regression best linear unbiased prediction-method 6 (rrBLUPm6) with well-known genomic selection methods (rrBLUP, GBLUP and BayesA) in terms of accuracy of prediction, computing time and memory requirement. The impact of the genetic architecture and heritability on the accuracy of genomic evaluation was alsostudied. To this end, a genome was simulated which consisted of five chromosomes, one Morgan each, on which 5000 biallelic single nucleotide polymorphisms (SNP) were distributed. Prediction of genomic breeding values was done in different scenarios of number of QTL (50 and 500 QTL), distribution of QTL effects (uniform, normal and gamma) and different heritability levels (0.1, 0.3 and 0.5).Pearson’s correlation between true and predicted genomic breeding values (rp,t) was used as the measure of prediction accuracy. Computingtime and memory requirement were also measured for studied methods. The accuracy of rrBLUPm6 was higher than GBLUP and rrBLUP, and was comparable with BayesA. In addition, regarding computing time and memory requirement, rrBLUPm6 outperformed other methods and ranked first. A significant increase in accuracy of prediction was observed following increase in heritability. However, the number and distribution of QTL effects did not affect the accuracy of prediction significantly. As rrBLUPm6 showed a great performance regarding accuracy of prediction, computing time and memory requirement, we recommend it for genomic selection.

The aim of this study was to compare the predictive performance of tree-based methods including regression tree (RT), random forest (RF) and Boosting (BT) in genomic selection. To do this, a genome comprised of five chromosomes was simulated for 1000 individuals on which 5000 single-nucleotide polymorphisms were evenly distributed. Comparison of methods was made in different scenarios of genetic architecture (number of QTL and distribution of QTL effects) and heritability level (0.1, 0.3 and 0.5). Computing time and memory requirement of the studied methods were also measured. In all the scenarios studied, the RT had the lowest accuracy, one-half to one-third of that was observed for RF and Boosting. Therefore, while RT was most efficient user of time and memory, because of its low accuracy, it was not recommended for genomic selection. Comparing RF and Boosting, at low levels of heritability (0.1 and 0.3), theprediction accuracy of RF was significantly higher than Boosting, but at heritability of 0.5, their accuracy was almost equal. In addition, RF was significantly superior to Boosting regarding computing time and memory requirement. While, heritability had a significant impact on the accuracy of prediction, the effect of number of QTL and distribution of QTL effects were not very dramatic. According to the overall performance of the studied methods, RF is recommended for genomic selection.

Owing to massive jump in DNA technology, large-scale genomic datasets, including valuable information, have become available. While this is a prodigious opportunity, and it can also be a big challenge because analysing these large datasets with current computers and software tools is very difficult and may take days or even weeks to complete. Novel approaches such as parallel computinghave been suggested to deal with these large datasets. Here, the effect of parallel computing on the performance of random forest (RF) algorithm for imputation of missing genotypes was studied. To this end, the genotypic matrices were simulated for, respectively, 500, 1000, 2000, and 3000 single-nucleotide polymorphism (SNP) for 500, 1000 and 2000 individuals, respectively. Then, 50% of genotypic information was masked and imputed by RF. The per cent of genotypes correctly imputed was used to measure accuracy of genotypeimputation. Serial and parallel computing were applied to the data. In comparison to serial computing, parallel computing did not affect the accuracy of imputation, and the accuracy was the same in both scenarios. However, regarding computational time, parallel computing accelerated the analyses significantly in a way that it reduced the running time up to 63%. This was due to the fact that in the serial computing, only 10% of the processing power of the central processing unit (CPU) of the machine was used by the RF, while in theparallel computing, 55% of the processing power of the CPU was utilized. Therefore, as parallel computing significantly reduced the computing time and does not affect the accuracy of the results, this approach should be exploited by researchers to analyse large genomic datasets.