Some empty functional annotations are reported by the API

[pverscha]

Sometimes, an empty functional annotation (in the format EC: or IPR:) is reported by the Unipept API. This causes downstream analysis of these samples to come up with weird or incomplete visualizations. After some initial investigation, I can see that these functional annotations are reported by the mpa/pept2data endpoint, so the issue is probably already the result from a parsing error when constructing the Unipept Database.

An example peptide for which this issue occurs is for example AAAALTER.

[pverscha]

I have been looking deeper into this issue and it seems that peptides for which none of the originating proteins report EC-numbers cause this issue.

Example peptide is: GGISVPGPMGPSGPR, with associated proteins:

• C0HLJ1
• C0HLJ5
• C0HLJ7
• C0HLH3
• C0HLJ9