Re: FW: Rejected posting to ASREML-L@AGRIC.NSW.GOV.AU

From: Bruce Southey <southey_at_UIUC.EDU>
Date: Wed, 23 Feb 2005 22:01:45 -0600

Well, sire models are also extremely biased under GLMM as there is 3/4 of the
genetic variance is unaccounted for (unless you assume normality and identity
link). The real problem is animal models is lack of information rather than the
algorithms per se (okay, some algorithms are plain bad). Like all analyzes, it
is user beware.


---- Original message ----
>Date: Thu, 24 Feb 2005 11:16:38 +1100
>From: Arthur Gilmour <arthur.gilmour_at_AGRIC.NSW.GOV.AU>
>Subject: Re: FW: Rejected posting to ASREML-L_at_AGRIC.NSW.GOV.AU
>Dear Bart,
>I tried to run a binomial model in ASREML, but got the surprising message
>that LogL is unsuitable for comparing GLMs. I know about the complexity of
>analysing binomial models, but I was unaware of the fact that 'the
>situation was that bad'. Is there any reference to this? And does this
>mean that it is no longer possible to run binomial models in ASREML?
>The LogL value reported by ASReml is for the analysis of the working
>As this working variable changes between iterations and especially between
>the LogL values are not comparable - even to confirm convergence of a
>particular model.
>ASReml also calculates a 'Variance heterogenity factor' which is the
>where Deviance is calculated for the underlying Maximum likelihooh model.
>In a GLM, (no random effects), this deviance can be used to compare
>However, for GLMM (with random effects), it does not represent the total
>(ignoring part related to the random effects). So it also is unsuitable
>as a basis
>for testing in GLMM models.
>Note that the GLMM procedures in ASReml are also seriously biassed when
>to Animal models (which is why they have not been described in the User
>Best wishes in the Name of Jesus Christ, my Saviour.
>Check out my 'CargoVale' website (see below)
>Arthur Gilmour PhD
>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;
>The ASReml discussion group
> has reactivated. To join,
>Movements: 6-11 Feb Biometrics meeting at Thredbo
> 14-18 Feb Armidale Summer School
>This message is intended for the addressee named and may contain
>confidential information. If you are not the intended recipient or
>received it in error, please delete the message and notify sender. Views
>expressed are those of the individual sender and are not necessarily the
>views of their organisation.
Received on Tue Feb 23 2005 - 22:01:45 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.)