Random primer sample doesn't check MIDI section of NS5b

[donkirkby]

In the 28-Jul-2016.M04401 run, sample 73053B-RP now notices that it is a single sample that covers both the whole genome and the MIDI ranges. Since the whole genome range has low coverage at the start, as well as NS3 and NS5a, the whole sample gets rejected. In version 7.9 on the old style of Kive run, it checked the MIDI section separately, and that had good coverage.

The MIDI section should have its coverage checked separately, even when there's a single sample that covers all of NS5b. This only applies when deciding whether to produce a report. Deciding to make a "susceptible" resistance call still requires coverage at all known resistance positions.