I'm wondering if it is possible to compare MN-ase and FAIRE (or DN-ase) results.
I tried to compare it using
bedtools intersect -v to count features shared and not-shared between files.
In theory, MN-ase gives opposite results to FAIRE or DN-ase. So great number of not shared features would mean that methods gives consistent results.
BUT, the coverage of MN-ase data is very low (0.53%) so it is very easy to hit with FAIRE peak in the interval not present in MN-ase data - so it inflate positive result.
Some features are covered by intervals from both files (ie. negative MN-ase and positive FAIRE), for such features I have run a t-test comparing length of overlap of intervals in each file. Still it doesn't seem valid.
Is there any way to compare such datasets?