Add heterozygosity test

[marcus1487]

New subcommand to identify regions with heterozygous bases. Should be some type of multimodal versus unimodal test. Will also have to filter out off-by-one regions, maybe by offset testing.

[marcus1487]

This test will likely have to come after a major overhaul in order to fix "off-by-one" (OBO) base alignment errors as these regions look very similar to heterozygous sites.

The OBO fix follows a seed and extend algorithm. This would proceed as follows:

1. Create squiggle reference
   • Median signal at each base across reads, excluding bases where too many reads do not fall within a particular range (determined statistically)
2. Identify stretches of reads where the signal matches better up or downstream one base for two or more consecutive bases
3. Extend those identified regions by as many bases as can confidently included in the off-by-one (OBO) region
4. For each such identified segment of a read, identify one segment to one side of the OBO region to remove and add one segment on the other side of the OBO region (as appropriate to shift the OBO region into phase with the squiggle reference).
5. Iterate.

This should produce a much cleaner overall alignment of raw signal which will allow heterozygosity tests and increase accuracy of modified base tests.