Don't report resistance if all regions have bad coverage

[donkirkby]

We recently changed the way MIDI samples get combined with their whole-genome partners, particularly in issue #462. Those changes caused some strange results, like sample 90154A_HCV in the 22-Mar-2019 run. It mapped nicely to HCV-1a, while its MIDI partner 90154AMIDI_MidHCV mapped nicely to HCV-1a and badly to HCV-3a. However, it now reports resistance for HCV-1a and HCV-3a.

Resistance should only be reported for a genotype if at least one of the four regions has good coverage:

• NS3
• NS5a
• NS5b (whole genome part)
• NS5b (MIDI part)

Currently, a region's coverage is considered good if it has average coverage of at least 100, and the start and end positions have coverage of at least 100.

Options for new coverage requirements:

• All positions have coverage of at least 100.
• All resistance mutation positions for that genotype have coverage of at least 100.

[donkirkby]

Chanson asked for the second option.