[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]


Hi Jeremy,

I think you can just fit a model like:

svol ~ mu test test.rep !r tree sca

and you'll get predictions for sca (and var(sca) = 1/4 var(dom)). I
would use a test.rep term to specify the nested nature of the
replicates. Now, given that probably you have a randomised complete
block design (the most typical design in Forestal Mininco), I would
treat blocks as random as well, so:

svol ~ mu test !r test.rep tree sca

I think that an option to include all material (OP and CP) in the same
analysis could be the use of genetic groups (!GROUPS in ASReml). I
don't see a problem in using an animal (or individual sounds better
with trees) model, specially because it will allow you to perform
forward selection.

The other point, is that with your current model you are not taking in
to account the heterogeneous variances for each site (you are fitting
a single error variance).

Good luck and let me know if you can run the program.

Best regards,


PS Regards to Victor Sierra.

Dr Luis A. Apiolaza
Quantitative Geneticist
CRC for Sustainable Production Forestry
School of Plant Science
University of Tasmania
GPO Box 252-55
Hobart TAS 7001

phone:   +61-3-6226 2213
fax:     +61-3-9229 2698

'There is no intellectual exercise which is not
ultimately useless'
Jorge Luis Borges

On 5/14/2001 you wrote:

Jeremy> Hello ASREMLers 
Jeremy>         I have a sort of unique problem that arises in tree breeding.  I
Jeremy> have included a subset of the data and the .as file.  The problem I have is
Jeremy> that alot of trees have been used as females and males and I cannot recode
Jeremy> them to make females always differ from the males.  
Jeremy> example ped file
Jeremy> 1 A B
Jeremy> 2 B C
Jeremy> 3 A C
Jeremy>         I would like to get the BLUP's for the parents.  The problem I have
Jeremy> is when I specify female and male in the model, I get a prediction for male
Jeremy> and female, and they are different.  So I have been forced to use an animal
Jeremy> model, results seem resonable but I would also like to get at the
Jeremy> interaction for each cross (Specific Combining Ability, a portion of the
Jeremy> dominance effects, or the deviation from the mean of the parents BLUPs...
Jeremy> however you would like to look at it).  I cant think of how to get at that
Jeremy> and have a unique BLUP for each parent.  I included a factor in the dataset
Jeremy> labled SCA that is unique for each cross, but since some of the individuals
Jeremy> come from open pollinated trials there are quite a few '.'s or unspecified
Jeremy> lables.  I would like to analyze the trials together, maybe someone has an
Jeremy> idea that will allow for this, or maybe I should just analyze the control
Jeremy> pollinated tests seperately from the open pollinated tests.
Jeremy> Any suggestions you may have as to how I should proceed would be greatly
Jeremy> appreciated.
Jeremy> Jeremy
Jeremy>  <<coast1.txt>>  <<>> 

Asreml mailinglist archive: