摘要
随着智慧畜牧研究的深入和高通量组学平台的发展,动物育种逐渐进入大数据时代的“育种3.0”,而机器学习是大数据研究不可或缺的有效手段。机器学习亦称计算机自动获取知识,是一门多领域交叉学科,涉及概率论、统计学等多门学科,作为人工智能和数据挖掘中的热门算法之一,具有学习能力强、准确率高和泛化能力强等优势,已经成为生物信息分析领域处理大型数据和进行预测的重要工具。目前广泛用于基因组、转录组、蛋白组、代谢组等多组学数据的整合与分析,尤其在家畜基因组估计育种值(genomic estimated breeding value,GEBV)的计算、基因型数据的填充和蛋白质结构及功能的预测等方面取得重要突破。作者介绍了几种常见的机器学习算法的概念和原理,对机器学习算法在重要家畜(猪、牛、羊)遗传育种方面取得的重要研究成果进行综述,进一步讨论了各种机器学习算法的优缺点、以及应用在动物遗传育种中存在的一些问题。最后,对机器学习的未来发展进行了归纳及展望,旨在提高预测准确性及效率,加快种群遗传进展,快速实现精准育种。
With the progress of research on intelligent animal husbandry and the development of high-throughput omics platforms,animal genetic breeding has been gradually entered the“Breeding 3.0”in the era of big data.Machine learning is an indispensable and effective means of big data research.Machine learning,the automatic acquisition of knowledge by computers,is a multidisciplinary cross-discipline,involving many disciplines such as probability theory and statistics.As one of the popular algorithms in artificial intelligence and data mining,it has many advantages such as high learning rate,good generalization ability and high accuracy.It has been an important tool in the field of bioinformatics analysis for processing big data and making predictions.Nowadays,machine learning is widely used for the integration and analysis of genomic,transcriptomic,proteomic,metabolomic and other multi-omics data,especially in the computing of genomic estimated breeding value(GEBV),genotype imputation,the prediction of protein structure and function of livestock,and other outstanding achievements.Firstly,the author introduced the concepts and principles of several common machine learning algorithms.Besides,the outstanding research results achieved by machine learning algorithms in genetic breeding of important livestock(pigs,cattle,sheep,and goats)were reviewed and further discussed the advantages and disadvantages of certain machine learning algorithms as well as some of the problems in animal genetic breeding.Finally,the future development of machine learning was summarized and prospected,aiming to improve the accuracy and efficiency of estimation,accelerate the genetic progress of populations,and rapidly implement precision breeding.
作者
周铂涵
梅步俊
吕琦
王志英
苏蕊
ZHOU Bohan;MEI Bujun;LYU Qi;WANG Zhiying;SU Rui(College of Animal Science,Inner Mongolia Agricultural University,Hohhot 010018,China;Agriculture Department,Hetao College,Bayannur 015000,China)
出处
《中国畜牧兽医》
CAS
CSCD
北大核心
2024年第12期5348-5358,共11页
China Animal Husbandry & Veterinary Medicine
基金
内蒙古自治区高等学校“青年科技英才支持计划”(NJYT22038)
青年教师科研能力提升基金项目(BR220112)
2030科技创新-重大项目(2023ZD0405102)
财政部和农业农村部:国家绒毛用羊产业技术体系(CARS-39)
内蒙古自治区高等学校创新团队发展计划项目(NMGIRT2322)。
关键词
机器学习
大数据
动物遗传育种
遗传改良
machine learning
big data
animal genetics and breeding
genetic improvement
作者简介
周铂涵,E-mail:1127078375@qq.com;通信作者:王志英,E-mail:wzhy0321@126.com;通信作者:苏蕊,E-mail:suruiyu@126.com。
