Small changes fixing PE coverage

I am trying to identify a minimal set of code changes that correct outputs in the statistics.csv file. I think it might be the following:

1. apply utilities.correctEdge() to paired_whole_coverage array inside of readsCoverage().

2. Make statistics.csv report _always_ write out paired_whole_coverage rather than whole_coverage. They should be equivalent in single-end case, and whole_coverage is simply wrong in PE case. So just use paired_whole_coverage always, don't bother testing for it.

Trivially, (2) might fix the problem by just making PT report the coverage that it is actually using, if the conditional there is causing a problem. But since the statistics.csv file is our source of truth, it looks very wrong. Maybe this will fix it.