猪饲料有效能值预测模型的构建

为探索饲料常规成分及碳水化合物组分与饲料有效能值之间的关系方程,本研究以NRC第11版《猪营养需要量》中发布的122套饲料营养成分表为基础,将饲料中11种基础成分[6项常规成分:干物质、粗蛋白质(CP)、粗纤维(CF)、粗脂肪(EE)、酸性醚提取物、粗灰分(ash);5项碳水化合物组分:淀粉(ST)、中性洗涤纤维(NDF)、酸性洗涤纤维(ADF)、半纤维素、酸性洗涤木质素]作为自变量,将饲料中的消化能(DE)、代谢能(ME)及净能(NE)作为因变量,采用SAS软件中的REG过程,分别建立不同性质饲料、自变量的不同组合与DE、ME及NE之间的回归关系方程,并以相关系数(R2)及变异系数(CV)作为评价回归模型的优劣。研究表明,有效能值与CP、ST及纤维类指标显著或极显著相关(P0.05或P0.01)。将所有饲料作为研究对象时,饲料的DE、ME及NE与上述11种基础成分之间建立的普适性回归模型预测效果较差。当将14种玉米及其加工产品形成子集时,建立饲料基础营养成分与DE、ME及NE的关系方程分别为7、6和7套(P0.05),且3组回归模型R2分别为0.632 8~0.772 3、0.646 9~0.684 9和0.670 5~0.822 1,CV分别为6.61%~8.40%、6.58%~7.34%和6.21%~8.27%;当将13种大豆及其加工产品形成子集时,共建立饲料基础成分与DE、ME关系方程分别有3和4套,回归模型R2分别为0.907 1~0.926 9、0.890 7~0.922 3,CV分别为5.40%~6.09%、5.79%~6.78%,NE与基础营养成分指标之间无法建立具有营养学意义的有效回归方程。对于同类饲料中具有相同自变量组合的DE及ME预测模型而言,两者之间的差异主要是自变量CP的系数上,且CP部分对ME的正效应低于DE,这保证模型预测的ME低于DE。同时选用本研究构建的适宜模型,补充了NRC第11版成分表中第97(去皮大豆粕,低寡糖,浸提)、101(全脂大豆,高蛋白质)及102号(全脂大豆,低寡糖)饲料的DE值分别为15.99、17.35、17.27 MJ/kg,ME值分别为14.53、16.15和16.14 MJ/kg。综上,以NRC(2012)饲料营养成分表为基础,建立的普适性有效能值回归模型预测效果较差。按照玉米类和大豆类进行分类,可建立DE、ME和NE与饲料化学成分之间的多元回归方程,其中最优的预测因子为CP、EE、ST、ash、NDF、ADF。具有相同自变量的同类饲料DE和ME预测模型之间的差异是CP系数,CP影响DE转化为ME的效率。     

Including an additional systematic environmental effect within a generation in an evaluation model improves accuracy of prediction of breeding values in a closed herd of pigs

The present study evaluated the advantage of mixed‐model techniques over a selection index under different magnitudes of an additional systematic environmental effect (ASEE) in terms of accuracy of prediction and expected genetic gain. The data attempted to simulate a closed herd in a pig breeding program. The base population (G₀) consisted of 10 males and 50 females. Six generations (G₀ to G₅) were selected by using a selection index of three traits without overlapping. Additional systematic environmental constants with four levels in a generation were assigned from a uniform distribution at different ranges. Breeding values of animals in the last generation (G₅) were estimated on the basis of an index of individual phenotype (SI‐U), SI‐U adjusted for ASEE using a least‐squares mean (SI‐A), best linear unbiased prediction using an animal model excluding ASEE (AM‐E), and an animal model including ASEE (AM‐I). Accuracy of prediction and expected genetic gain were larger by the animal model than by the selection index, even if heritability of the traits selected was high and ASEE was set to zero. When ASEE was zero, the accuracy of prediction and expected genetic gain given by SI‐U and AM‐I were similar to those given by SI‐A and AM‐E, respectively. However, the differences in accuracy and expected gain between SI‐U and AI‐A and between AM‐I and AM‐E increased as the range of ASEE increased. It was concluded that selection based on an animal model was more effective than index selection, even if the herd environment was uniform and traits with high heritability were selected, and that it should be always included in an evaluation model, however slight any systematic environmental effect may be in a closed herd.     

利用广义线性模型预测家畜离散性状育种值

将广义线性混合模型(GLMM)引入动物离散性状的遗传分析及个体的遗传评定,初步比较了GLMM方法与一般线性方法(LM)的估计效果。模拟研究的性状为单阈值二项分类性状,选用的连接函数为对数连接μi=eη/(1+eη),方差函数为V(μi)=μ(i1-μi)/n,试验设计为全同胞-半同胞混合家系,参数估计采用Fisher迹法。结果表明:GLMM方法能较准确地估计公畜的个体育种值,在个体的遗传评定效果方面要明显优于常规的线性方法,其预测的育种值排序结果与真实育种值的排序之间存在极显著的相关性(P<0.001)。     

反刍动物采食量的概念与研究方法

本研究对反刍家畜采食量的研究发展与现状进行了回顾与总结,采食量及其相关术语的定义主要基于饲草料或家畜,其测定方法和精确含义有差异;反刍家畜采食是由生物和非生物因素相互作用共同影响的复杂动态过程,采食量影响因子主要分为家畜的(胃肠蠕动力、选择性采食等)、饲草料的(粗纤维含量、抗营养成分等)和饲养条件的(草层结构、草地饮水点分布等)因素3个方面;采食量的测定方法有基于牧草测定的直接法和基于家畜测定的间接法,后者较多地应用于放牧家畜;预测方法有基于数学函数和生物学原理的模型法,基于消化试验的经验法和结合家畜属性的改进经验法。舍饲条件下可准确地测定和预测家畜的采食量,但对放牧家畜还没有精确的方法。通过测定与饲草水平密切相关的家畜尿液、血液或粪便中的代谢物水平将成为预测放牧家畜采食量的准确方法。     

未来气候变化条件下青藏高原多年冻土区潜在植被类型面积预估

在青藏高原多年冻土区根据490个植被调查样点数据和3种遥感数据集的27个变量,利用决策树分类模型,模拟出4种代表性浓度路径情景下10个气候系统模式在2050年和2070年的青藏高原高寒草地类型(高寒沼泽草甸、高寒草甸、高寒草原、高寒荒漠和裸地)潜在分布结果。同时为保证数据分析的一致性,利用数据对当前高原草地类型也进行了反演。结果表明:相比当前高寒草地分布面积,预计在2050年和2070年裸地和高寒草甸面积微弱减少,高寒草原和高寒荒漠面积在微弱增加,高寒沼泽草甸面积变化不明显。结果在4种代表性浓度路径情景下的表现基本一致,研究不仅可以为高寒草地气候变化研究提供植被类型相关的数据支持,还可以为青藏高原多年冻土区碳循环的探讨提供部分的方法和理论依据。     

3种不同方法对肉牛胴体性状预测能力的比较研究

本研究为了寻求一种对肉牛胴体性状预测准确性较高的方法,运用DPS数据处理系统和SAS软件比较偏最小二乘回归、GM(1,N)灰色系统和BP神经网络3种常用的预测模型对肉牛胴体性状的预测能力。选择肉牛7个宰前生长性状(体高、体长、胸围、腹围、管围、宰前活体质量、平均日增体质量),对2个重要的胴体性状(胴体质量和净肉质量)进行预测。结果表明:偏最小二乘回归方法在肉牛胴体性状预测方面准确性最高;GM(1,N)灰色系统和BP神经网络预测准确度偏低。本研究还将3种预测结果相结合,取其均值,大大提高了预测的准确性。这一研究将为肉牛生产实践提供一定的科学参考。     

Model averaging for genome‐enabled prediction with reproducing kernel Hilbert spaces: a case study with pig litter size and wheat yield

Predictive ability of yet‐to‐be observed litter size (pig) grain yield (wheat) records of several reproducing kernel Hilbert spaces (RKHS) regression models combining different number of Gaussian or t kernels was evaluated. Predictive performance was assessed as the average (over 50 replicates) predictive correlation in the testing set. Predictions from these models were combined using three different types of model averaging: (i) mean of predicted phenotypes obtained in each model, (ii) weighted average using mean squared error as weight or (iii) using the marginal likelihood as weight. (ii) and (iii) were obtained in a validation set with 5% of the data. Phenotypes consisted of 2598, 1604 and 1879 average litter size records from three commercial pig lines and wheat grain yield of 599 lines evaluated in four macro‐environments. SNPs from the PorcineSNP60 BeadChip and 1447 DArT markers were used as predictors for the pig and wheat data analyses, respectively. Gaussian and univariate t kernels led to same predictive performance. Multikernel RKHS regression models overcame shortcomings of single kernel models (increasing the predictive correlation of RKHS models by 0.05 where 3 Gaussian or t kernels were fitted in the RKHS models simultaneously). None of the proposed averaging strategies improved the predictive correlations attained with single models using multiple kernel fitting.     

“Arte et Labore”—A Blackburn Rovers fan's legacy in human complex trait genetics

Through his own research contributions on the modelling and genetic analysis of quantitative traits and through his former students and postdocs, Robin Thompson has indirectly left a major legacy in human genetics. In this short note, we highlight examples of the long‐lasting relevance and impact of Robin's work in human genetics. A lone early study of marker‐assisted selection developed many of the tools and approaches later exploited (often after reinvention) by the human genetics community in GWAS studies and for prediction. Furthermore, a particularly clear example of the pervasive impact of Robin's work is that REML has become the default method to estimate variance components and that genetic predictions exploiting linkage disequilibrium in the population are starting to become used in precision medicine applications.     

生长猪大米蛋白消化能、代谢能的评定及预测模型研究

本试验通过采集我国多个地区大米蛋白样品,分析大米蛋白常规成分,测定生长猪消化能(DE)、代谢能(ME),建立大米蛋白常规成分与生长猪DE、ME之间的关系。试验选用体重(33.0±1.3)kg的"杜×长×大"三元杂交健康去势公猪12头,采用2个6×6拉丁方试验设计,包括1种基础饲粮和11种大米蛋白替代15%基础饲粮的试验饲粮。采用全收粪法和套算法结合测定大米蛋白猪DE、ME,并将大米蛋白的常规成分与猪DE、ME进行相关和回归分析,建立大米蛋白猪DE、ME预测模型。结果表明,11种大米蛋白风干基础下猪DE为(18.13±1.12)M J/kg,M E为(16.44±1.59)M J/kg。由此得出,大米蛋白猪DE最佳预测模型(绝干基础)为DE=22.17-0.51NDF(R2=0.50,RSD=0.93),DE=18.58-0.49 CF+0.31 EE(R2=0.70,RSD=0.77);M E最佳预测模型(绝干基础)为ME=21.42-0.74 NDF(R2=0.52,RSD=1.30)。NDF为大米蛋白猪DE、M E最佳预测因子。     

Predictive modes of action of pesticides in uterine adenocarcinoma development in rats

Endometrial adenocarcinoma in the uterine corpus is a malignant cancer that occurs in menopausal women and aged rodents. Because of the similarities in pathogenesis and morphology of endometrial adenocarcinoma in rodents and humans, prediction of the modes of action (MOA) in uterine carcinogenesis is important for extrapolation of rodent data to humans. Three MOAs have been accepted as major pathways for uterine carcinogenesis in rodents: 1) estrogenic activity, 2) increased serum 17beta-estradiiol (E2) to progesterone (P4) ratio and 3) modulation of estrogen metabolism to produce 4-hydroxyestradiol via P450 induction. Inhibition of estrogen excretion and increased aromatase in situ in the tumor are also a potential pathway. Here, chemicals showing uterine carcinogenicity were chosen from approximately 300 pesticides evaluated in Japan within the past decade, and their mechanisms were predicted using parameters from mechanistic and toxicity studies. Seven pesticides increased uterine tumor formation in rats, and the pathways of 4 pesticides could be predicted based on various mechanistic studies. The MOAs of cyenopyrafen and benthiavalicarb-isopropyl were predicted to be modulation of estrogen metabolism, while those of pyriminobac-methyl and spirodiclofen were predicted to be increased E2 to P4 ratio. The driven pathways of metazosulfuron and isopyrazam could not be predicted using several mechanistic studies. No mechanistic studies have been reported for sedaxane, which has a chemical structure and toxicological profile similar to isopyrazam. Our results indicated that appropriate mechanistic studies are useful for mechanism prediction in risk assessment. From this analysis, a flowchart showing a decision tree for predictive MOAs in uterine carcinogenesis was proposed.