Adding flags from gnomad to CohortAlleleFrequency

[theferrit32]

Also adding a notebook. Notebook could use some setup documentation

[theferrit32]

Open to changing this if someone can think of a better name. My understanding is that this is the number of samples which had this variant and contributed to the FAF which were indicated as heterozygous but which gnomad thinks may actually be homozygous. There is some scoring method and those over 0.9 are included in this count. gnomAD displays this count in a warning text at the top left under the variant summary stats. I put it as an integer here because we want to represent the same count.

[ahwagner]

heterozygousSkewedAlleleCount

[ahwagner]

Just catching up here, and I know this was merged already, but I'm not certain about this proposed change to our schema. I had planned to review this with @larrybabb on Wednesday, but will comment here with my thoughts before we go too far with these changes.

ancillaryResults is a good place to put additional findings / results from the study; counts of homozygotes, meanDepth, popMaxFAF95 are all examples of this. I can see lowComplexityRegion being an ancillary result as well. heterozygousAlleleBalanceFlagged per @theferrit32's description sounds like it should be nested inside the popMaxFAF95 object, as should the qcFilters. I want us to be careful not to treat metadata as discrete ancillary findings. It is also possible that other ancillary findings (e.g. popMaxFAF99) may use different filters or have different counts for these fields.

I'll think on the name for heterozygousAlleleBalanceFlagged if this is still open for consideration.