Protein annotates as Gene in UI, though symbols are correct

[sharatisrani]

When a path intends to show a protein (as denoted by the icon/symbol), it annotates as "Gene." Would be easier on the user if it said "Protein."

[sierra-moxon]

from TAQA:

• the icon on the gene is a protein and gene symbol.
• from a user POV we expect this is a "protein" that is being affected, the protein of the gene Hebp1. But in our graph building, disambiguation of genes and proteins causes a lot of noise (both too many edges and too few - some sources only annotate to "gene" ids when they should mean, biologically, "protein" -- they "pre-conflate" these two concepts). Because of this, the query has to "know" to look for a protein or gene, based on the source and no one really has a good way of doing those queries effectively. So we decided that NN should do "conflation" -- it's configurable -- if you query by "gene" you get results for "genes" and "proteins" -- it eases the searchability.
• polypeptide, protein, protein isoform were what the users were looking for, translated this into a protein category in the TRAPI, went looking by protein ID or protein symbol (which is not necessarily the same as the gene id).

from the UI perspective - we need to explain this. chemists might be more interested in the difference.
we like the way it works now.
e.g. if the person asks for a protein, we de-conflate at the user level maybe? return the protein instead of the gene, even if the underlying data sources use genes? If it is a conflated entity, it will always choose the gene first. Arbitrarily de-conflating would be dangerous.

should be tagged as a low priority to "fix" - if it is easy, change the icon.
we might be testing with a more critical eye.