摘要
【目的】探讨逆传播人工神经网络(BPANN)算法用于预测分子标记育种值的可行性。【方法】采用高通量测序技术对284尾F1代凡纳滨对虾及其父母本进行特定长度扩增片段测序(SLAF-seq),随机取200尾对虾样品的数量性状基因座(QTL)基因型和体质量数据,构建BPANN预测模型,利用该模型分别对其余84尾凡纳滨对虾进行体质量性状预测。【结果】构建了1个高密度的单核苷酸多态性(SNP)遗传连锁图谱,鉴定出6个与体质量相关的QTL,对此QTL的BPANN育种值预测结果显示,育种值的平均误差为0.0320±0.0064,低于贝叶斯线性回归模型预测的平均误差值(0.0462±0.0056)。【结论】BPANN用于预测凡纳滨对虾分子标记育种值效果良好。
【Objective】To explore the feasibility of the back propagation artificial neural network(BPANN)algorithm for predicting the breeding value of molecular markers,【Method】High-throughput sequencing technology was used to perform specific length amplified fragment sequencing(SLAF-seq)on 284 F1 generation of Litopenaeus vannamei and their parents,and the QTL genotype and weight data of 200 shrimp samples were randomly selected to construct a BPANN prediction model.The model was used to respectively predict the weight traits of the remaining 84 shrimps.【Result】A high-density single nucleotide polymorphism(SNP)genetic linkage map was constructed,and 6 weight-related QTLs were identified,and used to predict breeding values by the BPANN.The average error of the breeding value predicted by the BPANN prediction model was 0.0320±0.0064,which was lower than the average error value of the Bayesian linear regression model(0.0462±0.0056).【Conclusion】The BPANN algorithm has a good effect on predicting the breeding value of molecular markers in L.vannamei.
作者
杨琼
刘青云
李强勇
彭敏
杨春玲
童艳梅
曾地刚
陈秀荔
陈晓汉
赵永贞
YANG Qiong;LIU Qing-yun;LI Qiang-yong;PENG Min;YANG Chun-ling;TONG Yan-mei;ZENG Di-gang;CHEN Xiu-li;CHEN Xiao-han;ZHAO Yong-zhen(Guangxi Key Laboratory of Aquatic Genetic Breeding and Healthy Aquaculture/Guangxi Academy of Fishery Sciences,Nanning 530021,China)
出处
《广东海洋大学学报》
CAS
北大核心
2022年第3期122-126,共5页
Journal of Guangdong Ocean University
基金
广西创新驱动发展专项资金项目(桂科AA17204080)
国家现代农业产业技术体系广西创新团队建设任务书(nycytxgxcxtd-14-01)
国家虾产业技术体系建设任务书(CARS-48)。
关键词
人工神经网络
凡纳滨对虾
分子标记
育种值
artificial neural network
Litopenaeus vannamei
molecular marker
breeding value
