Use stored VEP consequence for protein effect histogram filtering

[sallybg]

Note that this change still includes the ability to translate coding variants and calculate protein effect in the absence of HGVS p. strings and/or VEP consequence information. The intention is to eventually get rid of that fallback code once HGVS and VEP information has been fully generated for everything in the database and automated.
When testing, note that every coding variant in the score set must have either VEP information or an HGVS p. string available in order to show the protein effect options. Currently, the VEP script/job fails to fill out the functional consequence for no-change nucleotide variants. Incoming changes in the api will handle this, but for now you can manually add "no change" to your local database for these variants to view protein effect histograms for more score sets (e.g. urn:mavedb:00000013-a-1 requires one variant's VEP to be manually filled out to "no change" to work).

[bencap]

Looks good, thanks Sally! Would you be able to add the issue for what we talked about on Slack regarding the VEP categories and removing the inferred consequence logic when you get a chance?