Must all samples in an eQTL analysis come from unrelated individuals?

Question

[Egl]

An eQTL analysis would be a nice supplement to a paper I am involved with. However, we would benefit a lot from having more samples. How important is it that all samples are from unrelated individuals?

We currently have 4 RNA-seq samples per donor during a treatment (time course study). We probably have enough samples to detect strong eQTLs if we use all the samples (i.e. N donors X 4 timepoints). However, I am afraid that this reuse of genetic material, albeit in differentially treated cells, would invalidate the results we obtain. Does anybody have any experience with this? I would appreciate any comments.

[Fabio Marroni]

There's a nice review on this in Nature Reviews Genetics. I have no experience in eQTL mapping but: a) I think you are not allowed to use repeated measurements (RNA-seq) for each individual UNLESS you explicitly model the fact that is a repeated measurement. b) It is not a problem if you use related or unrelated individuals, but again, you have to model this. I guess there are software packages that can do that for you. Reading the review I pointed out would be a good starting point.