特克塞尔
限制最大似然
生物
动物科学
遗传相关
遗传力
方差分量
多元统计
统计
遗传变异
数学
遗传学
最大似然
基因
作者
Steven Janssens,W. Vandepitte
标识
DOI:10.1016/j.smallrumres.2003.10.008
摘要
Genetic parameters were estimated with REML for seven body measurements and 13 linearly scored traits in Belgian Bleu du Maine, Suffolk and Texel sheep. Data were collected from 1993 to 2000 on 756 Bleu du Maine, 1153 Suffolk and 5232 Texel yearling ewes. Measured traits included body weight, height, length, body depth, loin width, heart girth and cannon bone girth. Skin thickness, tail thickness, tail implantation, tail position, pelvis slope, topline and legs were scored from 1 to 9 with one point intervals. For muscularity of the gigot half points were allowed. Fatness (F) was also scored in the scheme (1–9 by one point intervals) but was used as an independent variable, rather than a trait. Only the scored traits were standardised within assessor to remove variance heterogeneity between persons. Significant non-genetic effects in the final model were flock and assessor×year (fixed) and two covariables, namely "fatness" and "age". The additive genetic effect was the only random effect. Co-variance components were estimated using REML in multivariate animal models. The full phenotypic and genetic covariance matrix was formed by combining results from part analyses and using the technique of bending. Heritabilities of measured traits were in the range of 0.26–0.57 and genetic correlations between these traits were high. Heritabilities for scored traits were between 0.04 and 0.43 with rather high values for gigot muscularity traits in all breeds. Genetic correlations between the scored traits were weak except between the gigot muscularity scores and between some of the leg traits. The pattern of correlations was not uniform between breeds. A simplification of the scheme, without loss of applicability to different breeds, seems only possible by removal of some of the measured traits.
科研通智能强力驱动
Strongly Powered by AbleSci AI