(no subject)

Dear Rob,
According to your structure,
 
devr long ~ Trait !r Tr.Sire
1 2 1
0 0 0
Trait 0 US
Tr.Sire 2
Trait 0 US
Sire 0 0
 
The G header line indicates two dimensions (highlighted in red above), the outer dimension which you started and left incomplete should be writen as fully i.e genetic covariance matrix between your two traits. This should come right below the line; Trait 0 US and its a 2* 2 matrix. Afterwhich now the Sire 0 0 line follows (inner dimension). Otherwise if no covariance between the traits is assumed, the 2 should be replaced with 1.
 
Kind regards
Chris'l


Rob Brooks <rob.brooks_at_UNSW.EDU.AU> wrote:
Dear Asremlers,

I have recently moved from my old DOS version of ASREML to the newest
release and am having trouble running a simple analysis in which sire is
the only random effect. Basically it is failing to read the G structure
line, and I am bemused as this format for the G structure has always worked
for me. I must say that the specification of appropriate G structures has
always been something I have found frustrating, and am hoping that one of
you might spot an obvious mistake. It is a bivariate analysis of the traits
"devr" and "long".

Anything you spot that you would suggest I do differently, please let me
know. Many thanks

Rob


devr long ~ Trait !r Tr.Sire
1 2 1
0 0 0
Trait 0 US
Tr.Sire 2
Trait 0 US
Sire 0 0
............................................................................
............................
School of Biological, Earth and Environmental Sciences
The University of New South Wales
Kensington, Sydney 2052
NSW, Australia
PH: +61-2-9385-2587 FAUX: +61-2-9385-1558
http://www.bees.unsw.edu.au/research/groups/brookslab/brookslab.html

                
---------------------------------
Yahoo! Messenger NEW - crystal clear PC to PC calling worldwide with voicemail
Received on Wed Sep 08 2005 - 09:20:23 EST