Re: heritability estimation for binary and quantitative traits

Dear Mahboob,

ASReml cannot fit two binary variables on the underlying scale in a bivariate analysis in a direct way. You probbaly could achieve such an analysis by reformating your data file from say

animal fix Bin1 Bin2

multivariate layout

to

animal fix TRAIT Bin

univariate layout (twice as many rows)

and then fitting the model

Bin !bin ~ TRAIT TRAIT.fix !r TRAIT.animal ide(animal)
0 0 1
TRAIT.animal 2
TRAIT 0 US !GP
0.1 0.01 0.1
animal 0 AINV

where ide(animal) will represent the error covariance

I do not know that anyone has actually done this kind of thing.

Again, I have written this assuming an animal model but such a model
is not recommended and is likely to fail.

While I can help with syntax, you must take responsability for the model you fit and any interpretation. In particular, people are still writing PhD's on the analysis of binary data so I can't begin to do justice to the subject in an email.

------------------------
Arthur Gilmour

Retired Principal Research Scientist (Biometrics)

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

Sent using Mail2Forum (http://www.mail2forum.com).

Read this topic online here:
http://www.vsni.co.uk/forum/viewtopic.php?p=670#670

-------------------- m2f --------------------
Received on Sat Jul 24 2009 - 14:03:55 EST