母猪产仔模型的选优与模型参数的遗传分析

Optimization of Litter Size Model and Genetic Analysis of Model Parameters in Sows

根据姜曲海猪产仔数表现出的生命旋回与逐胎波动两大特征,分别用泊松旋回模型、三次多项式模型、改进二次多项式模型Ⅰ和改进二次多项式模型Ⅱ4种数学模型对其窝产仔记录进行了拟合。从统计和生物意义两方面综合评价,改进二次多项式模型Ⅰ为产仔数的优选模型。单性状动物模型和DFREML算法被用来估计优选模型参数的遗传力。研究结果表明,产仔模型参数A、B和模型纯二次曲线顶点估计值的遗传力较产仔数的遗传力高,提示针对产仔模型参数的选择将比直接选择产仔数更为有效。

According to the phenotypic characteristics of life-cycle and fluctuation with parities, four mathematical models including Poisson cycle model, cubic polynomial model, modified quadratic polynomial model I and modified quadratic polynomial model Ⅱ were respectively used to fit the expected litter size at different parities for Jiangquhai sows. From the viewpoints of statistics and biological meaning, modified quadratic polynomial model I was the optimum model of litter size. Single trait animal model and DFREML procedures were involved to estimate the heritabilities for optimum model parameters. The results showed that the values of heritabilities for model coefficients A and B and the estimation of acmes of pure quadric curve in model were higher than that for litter size, which hinted that selection on model parameters should be more effective than direct selection on litter size.