Re: Issues with Binomial variables, missing values & workspa

From: Arthur <asremlforum_at_VSNI.CO.UK>
Date: Fri, 21 Aug 2009 10:57:53 +0100

Dear Tony,

1) The !DV* qualifer pertains to ASReml 3 and requests dropping if records where SEX is missing. In ASReml 2, use !D -9 [ which will dicard records with SEX missing or coded as -9).

2) !MVINCLUDE will code any missing values in the design factors as 0. In your case, that is either a missing value in TEMP or in ID. You said you have 'missing data in the PEDIGREE' and that you do not want to 'include the missing values'. If you mean that ID is coded missing for some records and that you want to delete those records from the analysis, then use !D-9 on the ID variable.

3) Start with a binary scale analysis. In fact, unless TEMP is having very strong effects on the Sex ratio, analysis on the 0,1 scale will probaly be optimal. Use of an animal model on binary data has generally yielded biassed estimates of variance parameters in simulation studies.

4) The pedigree problem probably implies a problem with the pedigree.
The !SORT qualifier attempts to sort the pedigree and produce a new
pedigree file called Example.Ped.SRT. There could a problem in this file.
Examine it visually and do a spot check.
If ASReml finds a problem, there will be messages in the .asr file indicatting errors (which may be too cryptic to understand). So, you could rename Example.Ped.SRT as say ExampleS.ped and see if it still needs sorting, if so use !SORT again.
The !SORT option in ASReml 3 is slightly more robust.

5) Once you have the analysis running without !BIN !LOGIT,
it should be straight forward to add these qualifiers. The only
problem I can anticipate is if Tempm were a factor and some levels
were entirely one sex or another.

I trust this helps.

Arthur Gilmour

Retired Principal Research Scientist (Biometrics)

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

Sent using Mail2Forum (

Read this topic online here:

-------------------- m2f --------------------
Received on Sat Aug 21 2009 - 10:57:53 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.)