Re: Binomial

Hi,
A few quick answers:

> 1) Version 1.10 supports generalized models? (that section in the
> 1.10 version manual is missing, but is in the manual for the beta
> version 1.62 and older versions manuals)

ASREML always has supported generalized linear models. But as you point out you
have to see the pre-version 1 manuals (although the latest manuals may have
it).

>
> 2) How to calculate the heritability and the s.e. from a binary trait
> analysis?

Just as you would a normally distributed trait. This is in the old manuals for
the logistic models. Basically you use the variance of the link function as the
residual variance.

>
> 3) Some good references regarding generalized mixed models,
> particularly for genetic analysis?

There are papers (like mine in J. Anim. Sci.) that use ASREML. However, a good
source is Steve Kachman's work:

http://statistics.unl.edu/faculty/steve/

(Note he will be given a course at Iowa State University this Summer on
ASREML.)

I would suggest starting on binary data and Collett's book on binary data
analysis is an excellent one to start with. Then work up from that.

Regards
Bruce
Received on Thu May 20 2005 - 15:22:47 EST