Level-biases in estimated breeding values due to the use of different SNP panels over time in ssGBLUP.

摘要: The main aim of single-step genomic predictions was to facilitate optimal selection in populations consisting both genotyped and non-genotyped individuals. However, spite intensive research, biases still occur, which make it difficult perform across groups animals. objective this study investigate whether incomplete genotype datasets with errors could be a potential source level-bias between animals on different single nucleotide polymorphism (SNP) panels predictions. Incomplete erroneous genotypes young caused breeding values Systematic noise or missing data for less than 1% the SNPs had substantial effects differences animals, chips. individuals were biased upward, magnitude up 0.8 genetic standard deviations, compared Similarly, small value added diagonal relationship matrix affected level average Cross-validation accuracies regression coefficients not sensitive these factors. Because, historically, SNP chips have been used genotyping parts population, fine-tuning imputation within handling are crucial reducing bias. Although all estimating present chip incompleteness some might lead level-biases values.