[R-sig-ME] Non-Parametric with Lme4

[Ackley Lane]

Dear R-sig-mixed-models,

Following the guidance of a colleague I am trying to get variance
components from a mixed model, these are useful in calculating
heritabilities and genetic correlations.

My problem is that the data are not normal for each of the three measured
variables. I have read that non-parametric mixed models can be done where
vectors are calculated so that residuals can be obtained.

Any help would be incredibly useful (even if to suggest a better place to
post this question),

Thanks for your time,

Ackley

PS
Attached is a data set identical to mine, except that the dependent
variable has been replaced with random data, the data has one fixed effect
(TREAT) two random variables (MALE and FEMALE), nested in BLOCK. The
dependent variable in this case is RAND.