Investigate conflicting strand values in normalized gene groups

[jsstevenson]

While merging records for normalized gene generation, it looks like there are about 300 cases where an incoming record has a different strand value than the rest of the group. According to scientists (@ahwagner), this is a bit strange and worth investigating.

I threw together a quick TSV file tracking instances of this: https://gist.github.com/jsstevenson/03c198862d39595ebe0d07f9d6b2efa2 (lots of horizontal scrolling needed)