I have a set of 10 inbred mouse strains, 20 per strain, 200 mice total.
Because each strain is inbred, they have the same genes (like twins or
clones). Thus variation within strain is environmental, between strain
is genetic. I want to remove environmental and genetic variation from
one measure that is associated with another. In this case I have a size
measure of a brain part that correlates, not surprisingly, with total
brain size. I want to remove variation associated with brain size from
the brain part measure.
I was wondering if this is a situation for xt models? With regression
the problem is that the animals within strain are cannot be considered
independent. With ANCOVA I cannot assume parallel lines among strains
for the relation between brain part and brain.
brain part = constant + brain + strain + (mouse nested in strain) + error