Hi All,
I am having difficulty to run a bivariate model with a
binomial (0,1) and normally distributed variable. Any
suggestions will be greatly appreciated.
Please have a look and make suggestions if you can. This is
the second time I am asking for help and nobody has
answered yet.
Fikret ISIK
I keep getting error messages as blow:
1) jwe0019i-u The program was terminated abnormally with
exception code EXCEPTION_ACCESS_VIOLATION
Invalid Binomial/Threshold proportion
At the end of the .asr file:
2) Warning: The LogL value is unsuitable for comparing GLM
models Assuming binomial data expressed as counts
I checked the ASREML Mailing archives for similar problems
and Dr.Arthur Gilmour made some suggestions to one of the
user.
I have tried several suggestions to overcome the problem:
1- Increased the memory allocation in the command line(-S4)
2-Run Univariate first for each trait and got initial
variances and
placed them in the model.
3-In the error variance matrix, I used !GPF qualifier to
fixed the error variance for the binomial variable. None
did help.
I included the Command file for your review.
Here is some information about the data and the system I am
using:
family =120
test sites (locations)=4
total number of observations: 14883
System: Windows 98, Memory=160 MB, Intel Pentium Celeron.
ASREML version:[13 Jun 2001]
COMMAND FILE
-----------------------
Age1 Vpine data, single tree plots
treeid !P # tree id in Pedigree and Data files
female
male
test 4 !I # there are 4 sites, recode
rep 35 # There are 35 reps
family 120 !A # there are 120 families, coded tree
ht
fork
c:\Vpine\Data\age1.ped !SKIP 1 !MAKE # pedigree file
c:\Vpine\Data\age1.dat -STEP !CONTINUE !ASUV #
ht fork !bin !LOG ~ Trait Trait.test Trait.test.rep ,
!r Trait.family Trait.family.test !f mv
1 2 2
0 0 0
Trait 0 US !GPF
0.78028 0.792 1.000
Trait.family 2
Trait 0 US !GP
0.102466 0.102 0.129718
family 0
Trait.family.test 2
Trait 0 US !GP
0.032325 0.028 0.026134
family.test 0
-------------------
· Fikret ISIK, Post-doc
· N.C. State University, Dept. of Forestry
· Campus Box 8002, Raleigh NC 27695, USA
· phone: (919)515-5029, fax:(919)-515-3169
=====
.Fikret ISIK, Ph.D.
.Population genetics, Forest Tree Improvement
.NC State University, Dept. of Forestry
Campus Box 8002, Raleigh NC 27695-8002,US
.http:/www4.ncsu.edu/~fisik
.fax:919-515-3169 phone:919-515-5029
__________________________________________________
Do You Yahoo!?
Make a great connection at Yahoo! Personals.
http://personals.yahoo.com
--
Asreml mailinglist archive: http://www.chiswick.anprod.csiro.au/lists/asreml