Using single-step genomic best linear unbiased prediction to improve the efficiency of genetic evaluation on body weight in Macrobrachium rosenbergii

文献类型： 外文期刊

第一作者： Liu, Junyu

作者： Liu, Junyu;Kong, Jie;Sui, Juan;Luo, Kun;Dai, Ping;Lu, Xia;Meng, Xianhong;Luan, Sheng;Liu, Junyu;Kong, Jie;Sui, Juan;Luo, Kun;Dai, Ping;Lu, Xia;Meng, Xianhong;Luan, Sheng;Yang, Guoliang;Tang, Qiongying;Yang, Guoliang;Xia, Zhenglong

作者机构：

关键词： ssGBLUP; pBLUP; body weight; Macrobrachium rosenbergii

期刊名称：AQUACULTURE （ 影响因子：4.242； 五年影响因子：4.723 ）

ISSN： 0044-8486

年卷期： 2020 年 528 卷

页码：

收录情况： SCI

摘要： In a shrimp selective breeding program, the genetic ties among founders or pedigree members are often unclear and could lead to biased estimates of the genetic parameters and predicted genetic gains using the conventional best linear unbiased prediction method based on pedigree (pBLUP). The accuracy of genetic evaluation can be improved by genomic selection with high-throughput molecular markers. This study was designed to estimate genetic parameters and predict genetic gains for body weight using single-step BLUP (ssGBLUP) with genotyped parents in the giant freshwater prawn Macrobrachium rosenbergii. The dataset was from 4 generations comprised 78,700 individuals, which were offspring of four founder populations. A total of 410 genotyped parents with 83,480 single nucleotide polymorphisms (SNPs) loci were obtained after quality control. Variance components were estimated using the animal model with the relationship matrix H combining SNP markers and pedigree data or the relationship matrix A from pedigree data. Prediction accuracies of estimated breeding value (EBV) and genomic EBV (GEBV) were calculated using five-fold cross-validation (CV), depending on the testing dataset from genotyped parents (CV1) or their ungenotyped offspring (CV2). The response to selection for body weight was evaluated by the difference of EBV or GEBV between generations. For 410 genotyped parents, correlation coefficient between lower triangular elements of the A and H matrices were 0.78 after correcting the pedigree. Heritabilities of body weight obtained using the pBLUP model with the A matrix and the ssGBLUP model with the H matrix were 0.11 +/- 0.02 and 0.25 +/- 0.03, respectively. For the CV1 scenario, the prediction accuracies of EBV and GEBV for body weight were 0.33 and 0.47, respectively; for the CV2 scenario, their accuracies were 0.24 and 0.28, respectively. The accuracy of GEBV for body weight predicted using ssGBLUP compared to pBLUP increased by 42.42% and 16.67% for the CV1 and CV2 scenarios, respectively. After performing 3 selections, the cumulative genetic gains of body weight for pBLUP and ssGBLUP were 4.46 g and 10.88 g, respectively. Response to selection for body weight was underestimated through pBLUP (11.45%) compared to ssGBLUP (27.93%). Our results suggest that ssGBLUP with small number of genotyped parents can improve efficiency of genetic evaluation for body weight using a more accurate relationship matrix compared to pBLUP in M. rosenbergii.

分类号：

  • 相关文献
作者其他论文 更多>>

微信小程序