protein_ids_at_locus() gives incorrect results

[aeschriefer]

This line should be querying CDS, not transcript

pyensembl/pyensembl/genome.py

Line 580 in 1231e96

feature="transcript",

[iskandr]

I never thought to interpret it that way! If the UTR of a protein is affected by a mutation I still want to know the protein ID that is affected...but I guess that's up to individual preference? At the very least I guess the document should clearly specify the behavior.