Future TODOs

[lgeistlinger]

• phenotype association: can R's own lm be used instead of plink (to avoid external dependencies)? Are results the same?
• synchronize approach for phenotype association and expression association
• can we make use of VariantExperiment?

[lgeistlinger]

• multiple testing correction for eQTL studies

[viniciushdasilva]

LMM with correction for family structure

[d1eone]

Now that you support linear regression in base R, I was wondering if you could add the option to do logistic regression on a binary phenotype using a generalized linear model (glm function).

Implementing logistic regression (glm) looks like a very quick addition to your .fitGetPvalue function. I know that easier said than done with the scoping you have in your package, but I think it would add a lot to CNVRanger!

[lgeistlinger]

Thank you for the suggestion @d1eone. I am wondering whether you have any thoughts on this @viniciushdasilva?