Chanson asked that we not fail a sample for a tiny amount of contamination, such as ten read pairs. We're going to try using the same rules for on-target and off-target scores. If the region was off-target, but it has coverage of at least 50 at all key positions, then it should fail the sample.