Hi everyone,
So I am a little confused with the application of q-values or more broadly, FDR correction (BH or BKY).
I have run multiple linear mixed effect models to assess the impact of medication on several brain regions and have a total of 245 p-values in a matrix.
Do I apply the q-value method to the entire matrix of q-values? or should I apply it by region? (lets say there are appox 30 p-values for each region)
Likewise, If i was to use the BH or BKY FDR correction method, would I apply this to the entire matrix or by region?
If i do decide to go with the q-value approach, do I have enough p-values to produce robust q-values?
Finally, which method is more robust overall? I dont want to be overly conservative but I also want to adequately control for type-1 errors.
Note, there is no problem with my r code, I understand how to apply each FDR procedure, I'm just finding it difficult to find oncrete infromation on whether to apply to the entire model (which I would likely do with the q-value approach considering it will likely underperform with such low amounts of p-values by region) or by region?
Any advice would be greatly appeciated!
