Re: Genetic groups
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Genetic groups



Dear Erwin,

I have looked at your sample job.  The groups analysis as
you had it was not working.  It is because I assumed there would be more
than 1 genetic group.  If you change !GROUPS 1 to !GROUPS 2,
it will work.  I have changed ASREML to allow a single genetic group
[in the next release].  
{{However, I have recently upgraded my Fortran compiler on the SUN
  and the graphics library is now not working so I am unsure when I can
  produce a new version for release.}}
  
So, I now get [from the group.sln file\  
  mu                              1           0.000       0.000
  animal               C                      4.700      0.4000
  animal               A                      4.369      0.5615
  animal               B                      5.031      0.5615
  animal               1                      4.096      0.5224
  animal               2                      4.944      0.5224
  animal               3                      4.520      0.5224
  animal               4                      4.732      0.5224
  animal               5                      3.884      0.5224
  animal               6                      4.499      0.5224
  animal               7                      5.347      0.5224
  animal               8                      4.923      0.5224
  animal               9                      5.347      0.5224
  animal               10                     4.711      0.5224
              Residual            1         -0.5958       4.096
              Residual            2          0.5562       4.944
              Residual            3         -0.1980E-01   4.520
              Residual            4          0.2682       4.732
              Residual            5         -0.8838       3.884
              Residual            6         -0.4986       4.499
              Residual            7          0.6534       5.347
              Residual            8          0.7740E-01   4.923
              Residual            9          0.6534       5.347
              Residual           10         -0.2106       4.711

Compared with  [nogroups.sln]
 mu                              1           4.700      0.4000
  animal               A                    -0.3312      0.6483
  animal               B                     0.3312      0.6483
  animal               1                    -0.6041      0.5700
  animal               2                     0.2438      0.5700
  animal               3                    -0.1802      0.5700
  animal               4                     0.3178E-01  0.5700
  animal               5                    -0.8161      0.5700
  animal               6                    -0.2014      0.5700
  animal               7                     0.6465      0.5700
  animal               8                     0.2226      0.5700
  animal               9                     0.6465      0.5700
  animal               10                    0.1061E-01  0.5700
              Residual            1         -0.5959       4.096
              Residual            2          0.5562       4.944
              Residual            3         -0.1981E-01   4.520
              Residual            4          0.2682       4.732
              Residual            5         -0.8839       3.884
              Residual            6         -0.4986       4.499
              Residual            7          0.6535       5.347
              Residual            8          0.7741E-01   4.923
              Residual            9          0.6535       5.347
              Residual           10         -0.2106       4.711

These results are entirely consistent and do not show the problem
that erwin reported (below).  
[Of course I had different results before I fixed the group count
but then the solutions were obviously different converging to a
very different solution.]

It is important to specify the number of genetic groups in the !GROUPS
qualifier [apparently an over estimate is no problem].  Maybe
you had an underestimate in your example.  If so, I expect
ASREML might give a warning message.

Arthur


~
~


> 
> Dear colleagues,
> 
> I tried to fit some large animals models including genetic groups and 
> compared them with models without genetic groups. I carefully checked that 
> all unknown animals in the pedigree file were assigned to a genetic group.
> 
> In the models with genetic groups I found very high SE for estimates for 
> the effects of animal (derived from *.sln files) compared to the estimated 
> genetic variance. As an example, I found a genetic variance of 0.66 whereas 
> the SE of 
> an animal was 1.472, which seems to be much too high.
> 
> What might be wrong here ?
> 
> To test the genetic groups I compared the results of a very small example 
> (10 animals, two sires, unknown dams) with and without genetic groups 
> (group.as vs. nogroup.as, see below). Also here I ended with different 
> results.
> In the no-group version I found an estimate for A of -0.3105 (SE 0.6207) 
> whereas it was -0.500 (SE 1.905) in the version with genetic groups. I 
> expected the same results when all unknown animals are assigned to the same 
> genetic group.
> I am not quite sure if the differences of these small exemples are caused by 
> the size of the data or by mistakes in the *.as files.
> 
> Best regards,
> 
> Erwin Koenen
> 
> 
> 
> NOGROUP.DAT / GROUP.DAT
> 
> 1 3.5
> 2 5.5
> 3 4.5
> 4 5.0
> 5 3.0
> 6 4.0
> 7 6.0
> 8 5.0
> 9 6.0
> 10 4.5
> 
> NOGROUP.PED
> A 0 0
> B 0 0
> 1 A 0
> 2 A 0
> 3 A 0
> 4 A 0
> 5 A 0
> 6 B 0
> 7 B 0
> 8 B 0
> 9 B 0
> 10 B 0
> 
> GROUP.PED
> C 0 0
> A C C
> B C C
> 1 A C
> 2 A C
> 3 A C
> 4 A C
> 5 A C
> 6 B C
> 7 B C
> 8 B C
> 9 B C
> 10 B C
> 
> NOGROUP.AS
> 
> WITHOUT GENETIC GROUPS
>  animal !P
>  y
> nogroup.ped !ALPHA !MAKE
> nogroup.dat !MAXITER=20
> y ~ mu !r animal
> 
> GROUP.AS
> 
> WITH GENETIC GROUPS
>  animal !P
>  y
> group.ped !ALPHA !GROUPS 1 !MAKE 
> group.dat !MAXITER=20
> y ~ mu !r animal
> 
> 
> --
> Asreml mailinglist archive: http://www.chiswick.anprod.csiro.au/lists/asreml


<><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><>
Arthur Gilmour PhD                  mailto:Arthur.Gilmour@agric.nsw.gov.au
Principal Research Scientist (Biometrics)            fax: <61> 2 6391 3899
NSW Agriculture                                           <61> 2 6391 3922
Orange Agricultural Institute             telephone work: <61> 2 6391 3815
Forest Rd, ORANGE, 2800, AUSTRALIA                  home: <61> 2 6362 0046

ASREML is still free by anonymous ftp from pub/aar on ftp.res.bbsrc.ac.uk
    or point your web browser at ftp://ftp.res.bbsrc.ac.uk/pub/aar/ 

To join the asreml discussion list, send the message  
     subscribe
mailto:asreml-request@chiswick.anprod.CSIRO.au

To send messages to the list, mailto:asreml@chiswick.anprod.CSIRO.au

Asreml list archive: http://www.chiswick.anprod.csiro.au/lists/asreml

                        <> <> <> <> <> <> <>
"Christ Jesus came into the world to save sinners"   I Timothy 1:15.
<><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><>






--
Asreml mailinglist archive: http://www.chiswick.anprod.csiro.au/lists/asreml