Re: heritability estimation

From: Arthur <asremlforum_at_VSNI.CO.UK>
Date: Thu, 18 Jun 2009 10:58:26 +0100

Dear Derek,
I don't follow your email.
"I have calculated the heritability of a trait from a half-sib design using the standard formula, the animal model using asreml in SPLUS and the "interactive version" of asreml." Are there 3 methods here (Sire model, Animal mode (Splus version) and Animal model (Standalone ASReml version) or just 2 (SPLUS and Standalone animal models)?
"The first two agree but the second only gives me a value that is one half of what it should be. " Since I don't understand the first sentence I do not know what agrees in the second. But in any case, how can the first 2 agree, when the second is 'half', and was is the 'it'? heritability?
It would help if you cut and pasted the variance components calculated from each run. Then I could see whether there are two or three, and what exactly is different.
The .pin file is correct for the standalone asreml and the model shown.
ASReml-r possibly returns var( t( Agamma, Sigma2) while ASReml standaline returns (.in the .vvp file) var(Avariance, Sigma2) where Avariance is Agamma*Sigma2
[ASReml uses first order apprximations for the rescaling)
From the .vvp file values, its is relatively easy to calculate var(Avariance+Sigma2) and derive the se of heritability then using expressions in the ASReml userguide. However, ASReml-r does not have builtin functions for this.

Arthur Gilmour

Retired Principal Research Scientist (Biometrics)

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

Sent using Mail2Forum (

Read this topic online here:

-------------------- m2f --------------------
Received on Thu Jun 18 2009 - 19:58:26 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.)