bivariate model with some complications

From: awils15 <asremlforum_at_VSNI.CO.UK>
Date: Tue, 28 Jul 2009 15:39:16 +0100

Hello all,
I was wondering if anyone could help with the following problem. I'd like to fit a bivariate animal model between two traits whre one trait (dominance) has repeated measures while the second trait (y) which has only one record per individual. To allow env covariance between the traits I set this up with a permanent environmnet effect on both traits and fixed VR for trait 2 to a v small with the remainder pushed into the pe term, i.e.,

dom y ~ Trait !r Trait.animal Trait.ide(animal)
1 2 2

Trait.. ide(animal) 2
Trait 0 US

Trait.. animal 2
Trait 0 US

This model runs fine, but the extra complication is that dominance is actually scored as a binary trait here (0/1) and I wanted to try fitting teh corresponding GLMM. Although I have been able to run some univariate models of dominance using !BIN I am having trouble with the R structure for the bivariate version. Specifically I've been trying to fit

dom y !BIN !DISP~ Trait !r Trait.animal Trait.ide(animal)

But the variance structures as specified above are obviously inappropriate. The suggestion in the .asr output is to use ASUV but I am a bit unclear as to how/why. The data is currently set up in multivariate format, with one line per observation of dominance on an individual. A column for the second trait y has a record only for the first observation of a given individual.

Any suggestions would be very welcome!


-------------------- m2f --------------------

Sent using Mail2Forum (

Read this topic online here:

-------------------- m2f --------------------
Received on Wed Jul 28 2009 - 15:39:16 EST

This webpage is part of the ASReml-l discussion list archives 2004-2010. More information on ASReml can be found at the VSN website. This discussion list is now deprecated - please use the VSN forum for discussion on ASReml. (These online archives were generated using the hypermail package.)