采用期望最大化算法的半滑舌鳎性逆转性状高效遗传解析  

Highly efficient genetic analysis of sex reversal traits in Halfsmooth Tongue Soles using Expectation Maximization algorithm

作　　者：宋禹昕 常中宇 高进 赵云峰 杨润清[2,5] 蒋丽[2,5] SONG Yu-xin;CHANG Zhong-yu;GAO Jing;ZHAO Yun-feng;YANG Run-qing;JIANG Li(Wuxi Fisheries College/Nanjing Agricultural University,Wuxi 214081,China;Biotechnology Research Center/Chinese Academy of Fishery Sciences,Beijing 100141,China;National Experimental Teaching Demonstration Center for Aquatic Science/Shanghai Ocean University,Shanghai 201306,China;Hainan Academy of Ocean and Fisheries Sciences,Haikou 571126,China;Key Laboratory of Aquatic Animal Genomics,Ministry of Agriculture and Rural Affairs/Beijing Key Laboratory of Fishery Biotechnology/Chinese Academy of Fishery Sciences,Beijing 100141,China)

机构地区：[1]南京农业大学无锡渔业学院,江苏无锡214081 [2]中国水产科学研究院,生物技术研究中心,北京100141 [3]上海海洋大学,水产科学国家级实验教学示范中心,上海201306 [4]海南省海洋与渔业科学院,海南海口571126 [5]中国水产科学研究院,农业农村部水生动物基因组学重点实验室,北京市渔业生物技术重点实验室,北京100141

出　　处：《山东农业大学学报（自然科学版）》2024年第4期531-539,共9页Journal of Shandong Agricultural University：Natural Science Edition

基　　金：国家自然科学基金面上项目(32372848);中国水产科学研究院水产生物遗传大数据研究与应用创新团队项目(2023TD26)。

摘　　要：为了解析半滑舌鳎(Cynoglossus semilaevis)性逆转性状的分子遗传作用机制,定位筛选可用于性控育种的分子标记或侯选基因,本研究提出了一种期望最大化算法(Expectation-Maximization algorithm,EM),并基于该算法开展了半滑舌鳎性逆转性状的全基因组关联分析。EM算法直接使用阈模型中隐含连续正态分布表型的期望作为因变量,用迭代最小二乘代替logit回归法的迭代重加权最小二乘,它具有比logit回归法更直观、更易于编程的优点。本研究采用显著主成分控制群体分层后,使用EM算法与logit回归对对半滑舌鳎数据进行GWAS(Genome-wide Association Study,GWAS)分析。结果显示,EM算法结果无明显的假阳性或假阴性,比logit回归法的检测效力更高。基于EM算法的全基因组关联分析共定位到13个与性逆转性状显著关联的QTN(quantitative trait nucleotide,QTN),其中3个QTN位于W染色体上,10个QTN位于Z染色体上。经过基因注释发现,上述定位获得的QTN位于LOC103396896、MALT1、ADGRD2、FBXl17、DMXl1、SMARCA2、DMRT1、LOC103397760、NEUR13和PDLIM5a基因区段内。当进行检索时发现,这些基因参与了其他物种中涉及性别决定或性腺发育等相关过程。本研究提供了一种基于EM算法的具有高检测效力的全基因组关联分析方法,同时也为半滑舌鳎的性逆转遗传机制解析和性控育种提供有效的理论指导。In order to elucidate the molecular genetic mechanism of sex reversal traits in Half-smooth Tongue Soles(Cynoglossus semilaevis)and to map molecular markers or functional candidates for sex control breeding,we proposed an expectation maximization(EM)algorithm and carried out genome-wide association analysis(GWAS)of sex reversal traits in Half-smooth Tongue Soles.This EM algorithm directly uses the expectation of continuous normal distribution implied by the threshold model as the dependent variable,and uses the least squares to replace the weighted least squares used by the logit regression,which has the advantage of being more intuitive and easier to program than the simplex algorithm.We introduce significant principal components to control population structure and EM algorithm and logit regression were used for GWAS of Tongue Sole data.The results showed that there were no significant false positives or false negatives by using EM algorithm,which was more powerful than logit regression in quantitative trait nucleotide(QTN)detection.A total of 13 QTN significantly associated with sex reversal traits were identified by GWAS based on EM algorithm,including 3 QTN on W chromosome and 10 QTN on Z chromosome.By gene annotation,it was found that the QTN obtained above were located in the gene regions of LOC103396896、MALT1、ADGRD2、FBXl17、DMXl1、SMARCA2、DMRT1、LOC103397760、NEUR13 and PDLIM5a,which contains sex determination or gonad development in other species.This finding not only confirmed the high power to detect QTN of EM algorithm,and it also provides effective theoretical guidance for the genetic mechanism of sex reversal and sex control breeding of Half-smooth Tongue Sole.

关 键 词：全基因组关联分析 半滑舌鳎 性逆转 主成分 期望最大化算法 广义线性模型 

分 类 号：S917.4[农业科学—水产科学]