Re: VCOV structure for GxE

Dear Jerko

My recommendation is first
to fit env as a fixed effect

second, to omit the gen main effect when fitting a structure to the
interaction since the genetic correlation in the structure
is the gen main effect.

Hence

yield !wt weight ~ mu env !r gen.env
1 1 1
0 !S2==1.0
gen.env 2
gen 0 GIV1
env 0 CORUV .5 1.

Now your previous runs should allow a better starting value where I have 0.5 and 1

You previously have V(giv(gen)) and V(gen.env)
replace 1 with V(giv(gen)) + V(gen.env)
and 0.5 with V(giv(gen)) /(V(giv(gen)) + V(gen.env))

Once this runs, you can try CORUH, FA1, and XFA models

I trust this gets you going.

There is a longer draft exposition of these I think at
ftp://asreml@cargovale.com.au:asreml@ftp.cargovale.com.au
in a file called something like CHasrxfa.pdf

i

------------------------
Arthur Gilmour

Retired Principal Research Scientist (Biometrics)

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

Sent using Mail2Forum (http://www.mail2forum.com).

Read this topic online here:
http://www.vsni.co.uk/forum/viewtopic.php?p=778#778

-------------------- m2f --------------------
Received on Wed Aug 25 2009 - 23:50:34 EST