Annotation Service needs to be extended to support multiple alternate alleles

[nalinigans]

This is something to keep in mind when we start using this functionality.

Originally posted by @jPleyte in #186 (comment)

[jPleyte]

When there are multiple alleles there end up being multiple ways to write the variant call, especially as the length of the ALT and REF sequence grows. Rather than looking comparing the multiple alternate alleles with the reference value, we should normalise them.

Karthic's note on this:

I'm assuming by match, you mean a simple string comparison, which (as you note) will not return any matches.
We face the same issue when combining multiple VCFs (combining samples each of which might be multi-allelic). We solve this in the same way you describe - normalizing the ALT alleles and then doing string matching. There's optimized code in GenomicsDB for normalizing and matching normalized alleles. It's tuned specifically for the combined VCF case, but if there are other uses I can try to make the interface to the module friendlier.