Negative interval coverage in HsMetrics output
I'm using Picard's CalculateHsMetrics to calculate statistics about my targeted sequencing project which includes 18 interval regions. For some reason I'm getting negative coverage for one of the intervals in 3 of my 39 samples. But this only occurs in my per-target output file--the per-sample file seems to be using a different value for calculating the average target coverage.
Can you give me any insight into why this might be happening? Are there some files I can attach to help investigate further?