Hello.
I'm writing about a possible issue when incorporating a phylogenetic
correlation structure (corPagel) in a linear mixed effect model (lme).
There are two techniques for incorporating a random effect in a linear
model:
t1 (it is the traditional way): m1 - lme( y ~ x, random =
Can you provide a data frame and tree which reproduces the problem? The
following code throws an error for both fit3 and fit4. I can't reproduce
your non-error! Also, you may want to update your nlme to the current
version (3.1-103).
library(ape)
library(nlme)
library(MASS)
set.seed(123)