Dear Sirs,

I have situation in which I query different sites on the genome (biologically independent) and for each site I get a measurement. I have a number of measurements that are close together on the genome and I call that a region. I then want to compare the measurements of a region between two biological specimens using a Mann Whitney U test (small number of measurements per region).

I want to stress that I am not repeatedly measuring the same thing, but that all measurements concern separate loci on the genome. We do know however that loci that are in close proximity to each other show strongly correlated measurements.

A reviewer pointed out to me that these measurements are not independent because they are correlated and that the subsequent test is therefore not valid.

In my opinion however there is no reason to assume that the value of one locus influences the value of the adjacent loci. Therefore the measurements are independent, regardless of their correlation.

On the internet I found various opinions. If measurements influence each other, they are dependent, that much is clear. But the term correlation also pops up a lot. Indeed there will be correlation in the case of measurements influencing each other, but if that is not the case and there is still correlation present, would you agree that the measurements are independent regardless of their correlation?

Thanks for any insights. Kind regards,