Dear Bjorn,
You can consider each site as a different trait, so running a multivariate
analysis you could estimate the genetic correlation between traits. This
correlation would be an estimate of g x e interaction. Concerning including
dominance, I think you can include a code in the model representing a
full-sib family, and the variance of that would give you an estimate of
dominance. I guess a program including 2 sites and dominance would look
like:
Analysis of 2 sites
tree !P
mother 10
father 20
Fsibfam 200
fixed effects...
vol1 !m0
Vol2 !m0
pedifile
datafile
vol1 Vol2~Trait Tr.fixed ... !r Tr.tree Tr.Fsibfam
2 1 2
nsite1 0 !s2=ve1
nsite2 0 !s2=ve2
Tr.tree 2
Tr 0 US va1 cova12 va2
tree
Tr.Fsibfam 2
Tr 0 US vd1 covd12 vd2
Fsibfam
!end
Maybe I'm wrong, but I thing the program would be something like that. I
hope this helps.
Cheers,
Luis
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Luis A. Apiolaza
Institute of Veterinary, Animal and Biomedical Sciences
Massey University
Palmerston North
New Zealand
mail: l.a.apiolaza@massey.ac.nz
www: http://www.crosswinds.net/~uncronopio
'Quote me as saying I was misquoted' - Groucho Marx
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
----Original Message Follows----
From: Bjorn.Hannrup@skogforsk.se
To: asreml@chiswick.anprod.csiro.au
Subject: Request
Date: Wed, 18 Aug 1999 12:21:22 +0200
Dear ASREML-users,
At our organisation, SkogForsk, we are working with forest tree breeding. We
have started to think of using ASREML for specific purposes in our breeding
evaluation. However, we are uncertain whether this analysis is possible to
perform in ASREML. We want to do the following ;
Estimate genetic correlations across trials where we have the same genetic
entries replicated, i.e. we have replicated progeny trials where
measurements are done on trees belonging to the same families but growing in
different environments. So far we have understood that this is not a problem
with ASREML. The problem arises from the special "tree case". Our families
are in many cases fullsib-families from a partial diallel mating design and
consequently we want to include dominance effects in the statistical model.
Is this possible to do in ASREML ?
Grateful for replies
Björn Hannrup
____________________________________________________
SkogForsk, The Forestry Research Institute of Sweden
http://www.skogforsk.se
Björn Hannrup, Dr (Forestry)
phone: +46 18 18 85 39, fax: +46 18 18 86 00
e-mail: bjorn.hannrup@skogforsk.se
___________________________________________________
--
Asreml mailinglist archive: http://www.chiswick.anprod.csiro.au/lists/asreml
______________________________________________________
Get Your Private, Free Email at http://www.hotmail.com
--
Asreml mailinglist archive: http://www.chiswick.anprod.csiro.au/lists/asreml