Permanent environment effects in covariance analysis

From: Celine Teplitsky <teplitsky_at_MNHN.FR>
Date: Thu, 23 Apr 2009 15:24:25 +0200

Dear all,

I'm trying to calculate genetic covariances for traits for which I have
repeated measurements. For some of those traits, the permanent
environment gets bounded to zero in univariate analysis, and in the
multivariate analysis if I don't fix PE to 0 for these traits, then the
model will not converge.
Hence my question is: for the PE, what make more sense: is it (1) to
keep the covariance between PE free to take any values because this
reflects the repeated measurements in the data or (2) set the covariance
to 0 because covariance between a variance and a null component does not
really make sense?

The choice in this quite affects the results, and I can't really get
which is the proper line of reasoning.

Many thanks in advance for any help,

All the best

Received on Fri Apr 23 2009 - 15:24:25 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.)