From a linear model like `mod <- lm(y ~ x)`, I can obtain a data frame
with all the variables involved in the model (in that case, `y` and `x`).
How could I get a similar data frame from an lme object, e.g. fitted as
`mod <- lme(y ~ x, random=~1|g)` ?
I know that `getData` might work if the variables come from a data
frame, but the result is `NULL` if it is not like that. And if there is
a transformation (e.g. `lme(log(y)~x, ...)`), the result of `getData`
gives the original, untransformed variable.