[GenABEL-dev] R2-polygenic questions on GenABEL forum

[Yurii Aulchenko]

Dear All,

there is an interesting question about polygenic models on GenABEL forum

http://forum.genabel.org/viewtopic.php?f=6&t=34

Even leaving 'multiple...' part apart, I wonder if notion of R2 is any
good for polygenic model. Would something like

1 - var1/var0

where var0 is the total phenotypic variance, and var1 is 'residual'
variance when both fixed effects and estimated breeding values are
subtracted be fulfilling the definition?

Anyways, I have little idea how to answer that question, and wonder if
one of you has an answer, and could do that at the forum. I would be
most interested to learn!

best wishes,
Yurii