On the PCR test, which is completely bogus because it is run at 35-40X (logarithmically) and is unreliable after 30X.
I agree that too many repeats causes some false results.
How many repeats are being done currently?
And is the single marker for O affected by the number of repeats?