Investigate why certain metadata blurbs have chromosomal identifiers in their json blobs

[bencap]

If you run the below SQL query, you'll observe some score sets in this family contain metadata blobs that look like

{"protein": {"sequence_id": "ga4gh:SQ.Ya6Rs7DHhDeg7YaOSg1EoNi3U_nQ9SvO", "sequence_type": "dna", "sequence_accessions": ["NC_000001.11"]}}

This is incorrect, given that the protein blob should not contain an NC accession.

SELECT
	urn,
        mapping_state,
	post_mapped_metadata
FROM
	scoresets
	JOIN target_genes ON scoresets.id = target_genes.scoreset_id
WHERE
	scoresets.urn LIKE 'urn:mavedb:00001214-%'

[bencap]

It seems very likely this has something to do with all variants in these score sets failing to map.

[bencap]

Note that I've since manually updated these so that we could ingest uniprot IDs without a job failure, but if we were to re-run mapping this would occur again.