不同时期牙鲆形态性状对体重影响的通径分析及曲线拟合研究  

Path analysis and curve estimates of morphometric traits and body weight of Paralichthys olivaceus at different growth stages

作　　者：陈红林[1,2] 田永胜[1] 刘峰[1,2] 李洪[1,2] 李祥孔 宋莉妮 陈松林[1] 

机构地区：[1]中国水产科学研究院黄海水产研究所,青岛海洋科学与技术国家实验室、海洋渔业科学与食品产出过程功能实验室、农业部海洋渔业可持续发展重点实验室,山东青岛266071 [2]上海海洋大学水产与生命学院,上海201306 [3]大连海洋大学水产与生命学院,辽宁大连116023

出　　处：《中国水产科学》2016年第1期64-76,共13页Journal of Fishery Sciences of China

基　　金：国家863计划项目(2012AA10A408);国家自然科学基金项目(31372510;30972244);山东省泰山学者建设工程专项资助

摘　　要：以8月龄与14月龄的养殖牙鲆(Paralichthys olivaceus)为研究对象,分别测定体重(Y)和22个形态性状,对各性状进行相关分析,并建立多元回归方程,对各形态性状对体重的影响效果进行通径分析。分别以进入回归方程的各形态性状作为自变量,体重为因变量进行曲线模型拟合,筛选最优拟合模型。结果显示:(1)在不同生长阶段,相同形态性状与体重的相关性存在差异。(2)8月龄阶段,X_(18)(腹鳍基部到背鳍终点的直线距离)、X_4(体宽)和X_(11)(背鳍起点到臀鳍起点的直线距离)对体重的通径系数达到极显著水平(P<0.01),14月龄阶段,X_(18),X_(14)(背鳍终点到臀鳍终点的直线距离)和X_9(尾柄长)对体重的通径系数达到显著水平(P<0.05),由此可知,不同生长阶段,影响体重的形态性状不尽相同。8月龄和14月龄牙鲆形态性状对体重的多元回归方程分别为:Y=-119.541+7.191X_(18)+10.135X_4+7.197X_(11);Y=-484.931+31.959X_(18)+81.928X_(14)-17.899X_9。(3)8月龄阶段,将回归模型中的3个自变量分别与体重进行模型拟合,最优拟合模型均为线性模型,分别为:Y=-117.866+15.724X_(18);Y=-94.579+24.763X_4;Y=-100.602+33.184X_(11);14月龄阶段,将回归模型中的3个形态性状分别与体重进行模型拟合,最优拟合模型均为幂函数模型,分别为:Y=0.036X_(18)^(3.063);Y=0.095X_(14)^(2.587);Y=62.249X_9^(1.584)。本次研究表明,在牙鲆的不同生长时期,影响体重的主要形态性状不同,两阶段适用的最优拟合模型也不同,这为牙鲆不同时期的选择育种工作提供了理论依据。Paralichthys olivaceus is one of the most important aquaculture species in China. However, intensive breed- ing and many generations of inbreeding have resulted in poor disease resistance and slow growth. Therefore, it is nec- essary to improve the quality and productivity of P. olivaceus through selective breeding. To study the relationship be- tween morphometric traits and body weight ofP olivaceus at different growth stages, 83 fish were selected randomly （8 months old, n = 52 and 14 months old, n = 31） to measure 22 morphometric traits and body weights. The data were sta- tistically processed using correlation, path, and stepwise regression analyses, and multivariate regression equations were calculated. The traits were included in the equation as independent variables, and body weight was the dependent vari- able in six curve-fitting models to select the optimal model. The results showed that： （1）the correlation coefficients be- tween body weight and each trait varied and differed at the two developmental stages. （2）The path coefficients of X18, X4, and Xll for 8-month body weight were highly significant （P〈0.01）. The path coefficients of X18, X14, and X9 for 14-month body weight were also significant （P〈0.05）. The multivariate regression equation was Y = -119.541 ＋ 7.191 ）（18 ＋ 10.135 X4 ＋ 7.197 X11 for 8 months, and Y= -484.931 ＋ 31.959 ）（18 ＋ 81.928 X14 - 17.889 ）（9 for 14 months. This result suggests that the vital morphometric traits that affected body weight at the two growth stages were different. （3） All optimal models at 8 months were linear, including Y = -117.866 ＋ 15.724 X18, Y= -94.579 ＋ 24.763 ）（4, and Y= -100.602 ＋ 33.184 XI1. However, all optimal models at 14 months were power models, including Y = 0.036 X18^3.063, Y = 0.095 X14^2.507, and Y = 62.249 X9^1.584. These results indicate that the growth rhythms between the two growth stages were different. The R2 values for the multivariate regression equations at 8 and 14 months were 0

关 键 词：牙鲆 相关分析 通径分析 回归方程 模型拟合 

分 类 号：S917[农业科学—水产科学]