Re: How to use predict statement with existence of genetic group?

From: <arthur.gilmour_at_DPI.NSW.GOV.AU>
Date: Mon, 16 Jun 2008 10:24:05 +1000

Dear Peter,

I used the file "" from ASReml package as an example.
Appending a line "predict LINES" did not produce predicted values in
the .pvs file. After removing the genetic groups from the pedigree
file, ASReml gave me valid predicted values for LINES.

 animal !P
 sire 9 !A
ex11g.ped !ALPHA !MAKE !GROUP 3 !NEW
# ex11g_nogroup.ped !ALPHA !MAKE !NEW
AD ~ mu LINES !r animal 02.5 !GU
predict LINES

Is there any way to do prediction when genetic groups are specified?

The genetic groups in this case are the lines.
The genetic groups are fixed effects fitted within the 'random' animal
Consequently, if you examine the .sln file, you will see that mu and
LINES are singular.
(I fitted the equivalent SIRE model and got)

  LINES 1 0.000 0.000
  LINES 2 0.000 0.000
  LINES 3 0.000 0.000
  mu 1 0.000 0.000
  sire G1 176.9 3.942
  sire G2 162.7 4.738
  sire G3 183.5 3.377
  sire SIRE_1 177.0 3.537
  sire SIRE_2 180.1 3.537
  sire SIRE_3 173.5 4.147
  sire SIRE_4 157.7 3.674
  sire SIRE_5 167.6 3.864
  sire SIRE_6 183.3 3.804
  sire SIRE_7 186.9 3.453
  sire SIRE_8 184.2 3.616
  sire SIRE_9 179.7 3.453

since the SPARSEly fitted terms are fitted BEFORE the DENSE terms (mu

So, to predict LINE means, these are the GROUP means G1 G2 G3

PREDICT animal 1 2 3
should do the trick.

However, when I did it, (with the sire model) it gave the correct answers
but said they were 'aliased'!!

From the model
  mu LINE !r animal
predict LINE
I got
 LINES Predicted_Value Standard_Error Ecode
       1.0000 176.8961 3.9425 E
       2.0000 162.6613 4.7384 E
       3.0000 183.5276 3.3770 E
 SED: Overall Standard Error of Difference 5.739

From the model
  mu LINE !r sire # where sire has pedigree incorporating genetic groups
predict sire 1 2 3 !PRINT

Warning: 3 non-estimable [aliased] cell(s) may be omitted from the
                The Overall SED statistic includes non-estimable

 sire Predicted_Value Standard_Error Ecode
 G1 176.8961 3.9425 *
 G2 162.6613 4.7384 *
 G3 183.5276 3.3770 *
 SED: Overall Standard Error of Difference 5.739

This is evidently a problem with the algorithm for detecting ALIASSED
when GROUP effects are included in the animal BLUPS.

If I drop the redundent (singular) terms mu and LINE from the model,
the predictions are unchanged but no longer reported as aliassed.

AN interesting exercise. I trust this is helpful.

May Jesus Christ be gracious to you in 2008,

Arthur Gilmour, His servant .
Mixed model regression mapping for QTL detection in experimental crosses.
Computational Statistics and Data Analysis 51:3749-3764 at

Personal website:

Skype: arthur.gilmour,
Principal Research Scientist (Biometrics)
NSW Department of Primary Industries
Orange Agricultural Institute, Forest Rd, ORANGE, 2800, AUSTRALIA
fax: 02 6391 3899; 02 6391 3922 Australia +61
telephone work: 02 6391 3815; home: 02 6364 3288; mobile: 0438 251 426

ASREML 2 is now available from
The ASReml discussion group is at
To join it,
Archives are at

The ASReml3 alpha download registration page is:
There are asreml3alpha forums at:

Proposed travel:
Towoomba (MFX) 19-23 May
Melbourne Aust Stat Soc 3rd July
Argentina 9-19 September ASReml WS + Biometrics 
Received on Wed Jun 16 2008 - 10:24:05 EST

This webpage is part of the ASReml-l discussion list archives 2004-2010. More information on ASReml can be found at the VSN website. This discussion list is now deprecated - please use the VSN forum for discussion on ASReml. (These online archives were generated using the hypermail package.)