|
| 基于全基因组选择的长牡蛎肥满度分布参数预测方法 |
| A Method for Predicting Distribution Parameters of Condition Index of Crassostrea gigas Based on Genomic Selection |
| 投稿时间：2019-07-16 修订日期：2019-09-12 |
| DOI： |
| 中文关键词: 全基因组选择 单核苷酸多态性 二次特征选择 高斯通用加性模型 长牡蛎 |
| 英文关键词: Genomic Selection SNP two-stepwise selection Gauss generalized additive models Crassostrea gigas |
| 基金项目:国家自然科学基金项目（面上项目，重点项目，重大项目） |
|
| 摘要点击次数: 248 |
| 全文下载次数: 0 |
| 中文摘要: |
| 全基因组选择是一种用于改良动植物育种群体中数量性状的方法，通过使用覆盖整个基因组的分子标记信息对复杂性状进行预测，从而帮助筛选出更适合培育的亲本。本文基于长牡蛎的单核苷酸多态性（SNP）位点信息，提出了一种预测长牡蛎肥满度分布参数的全基因组选择的新方法。首先，采用一种基于不同评价准则的二次特征选择方法，挑选与目标性状相关性较高的SNP位点；其次，利用所挑选的位点信息构建具有正则化项的高斯通用加性模型对每个长牡蛎样本肥满度及其分布参数进行预测；最后，在长牡蛎数据上将本文方法和一些现有方法进行了验证比较。实验结果表明，本文所提出的方法具有更好的拟合精度和更低的均方误差，并能对样本性状稳定性进行有效地评估。 |
| 英文摘要: |
| Genomic Selection (GS) is a new method for improving quantitative traits in animal and plant breeding. By using genetic markers covering the whole genome of the species to predict complex traits, it can help screen out parents that are more suitable for breeding. Based on the single nucleotide polymorphism (SNP) locus information of Crassostrea gigas, this paper proposes a novel GS method for predicting Crassostrea gigas Condition Index. Firstly, a two-stepwise feature selection method based on different evaluation criteria is used to select SNP loci tightly bound to Condition Index. Secondly, the selected SNP is used to construct a generalized additive model under Gauss distribution with regularization terms for each sample. Finally, the method is compared with other methods by employing the Crassostrea gigas data. The results show that the proposed approach has better fitting accuracy and more accurate estimation variance. Meanwhile it can effectively evaluate the stability of sample traits. |
|
View Fulltext
查看/发表评论 下载PDF阅读器 |
| 关闭 |