The maternal environment effect var

The maternal environment effect variance component estimate is 1.59, the maternal genetic effect variance component is 1.444, and the direct genetic effect variance is 0.104. You will notice something strange about this result: the correlation between direct and maternal genetic effects is -1.502, which is out of theoretical bounds for a correlation coefficient. This can occur if we do not constrain the unstructured variance-covariance matrix for these two effects (we used the !GUUU qualifier on the model term). If we try to constrain the model in this case, it does not converge. This result suggests that we do not have a reliable estimate of one or more of the
variance-covariance terms in this model. Notice that the direct genetic effect variance is close to zero (and is smaller than its standard error). If a variance component really is zero, it cannot have a covariance with any other term. So, the trouble is probably occurring because this direct genetic variance component is approaching zero. We could simplify the model by dropping the direct genetic effects and compare the AIC values of the reduced model to this one to see if the reduced model could be accepted.

結果 (中文) 1:

孕产妇环境效应方差分量估计是 1.59，母体遗传效应方差分量是 1.444，直接遗传效应方差是 0.104。你会注意到了这个结果奇怪的事情︰ 直接和母体遗传效应之间的相关性是-1.502，超出理论界限的相关系数。如果我们不限制这两种效应的非结构化方差-协方差矩阵，这可以发生 (我们使用 ！GUUU 限定符对模型一词）。如果我们试图在这种情况下的约束模型，它不收敛。这就表明我们没有可靠的估计的一个或多个在此模型中的方差-协方差条件。请注意，直接遗传效应方差是接近于零 （和小于它的标准误差）。如果方差分量真的是零，它不能有任何其他术语与协方差。所以，麻烦可能会发生，因为这直接遗传方差分量几近于零。我们可以通过删除直接遗传效应简化模型，并比较这一查看如果能接受降阶的模型降阶模型的 AIC 值。