[R-sig-ME] lmer within family mixed model question.

[David Duffy]

On Mon, 25 Feb 2013, Sai wrote:

> Dear all,
>
> I am currently using lmer from lme4 package to check for family effect for
> a phenotype. Since Family has a genetic as well as environment component, I
> would like to look for environment component. For this I would like to look
> for phenotypic differences within siblings (which share 50% of genes from
> common parents). For example
>
>
> MourseID    Phenotype    Cage    Family
> 1    2.1    A       F1
> 2    4.0    A    F1
> 3. 1.0    B    F1
> 4    1.3    C    F2
> 5    1.9    C    F2
> 6    4.3    D    F2
> 7    2.2    D    F2
> 8    3.4    E    F2
> 9    3.5    E    F2
>
> Model1 <- lmer(Phenotype ~ 1+ (1|Cage)+(1+Cage/Family)
>
> The model would treat Cage as a separate random variable and Family nested
> within Cage (to check for difference between same family from different
> cages). I do not know how to formulate the above model to find the
> difference within F1 and F2 family mice?

So F1 and F2 are inter and back crosses?  Something like

Model1 <- lmer(Phenotype ~ 1+ (1|Cage)+(1|Family)) 
Model2 <- lmer(Phenotype ~ 1+ (1|Cage)+(1|Family)+(1|Family=="F2"))