Do we want more dev work on templates, for current creative-modes?

[colleenXu]

We narrowed the scope of #461 to #461 (comment) (with adjustments #461 (comment)).

However, there may be other work that we want to do, regarding the "treats" and "chem affects gene" templates:

Both

• Review potential templates added, but not used, in Template revisions bte_trapi_query_graph_handler#167 and Template revised (fixed merge conflicts) bte_trapi_query_graph_handler#169 (comment). These templates were based on my skimming of x-bte annotation, looking for potentially useful MetaEdges and data.
• Review KPs from other teams, and look into what MetaEdges we can use for creative-mode. I did a quick skim for Multiomics/Text-Mining (part of the effort above to add templates), but didn't fully understand the data available.
• using gene homolog/ortholog data. Related to this (those MetaEdges uses related_to since I wasn't sure what predicate to use), this not-done issue

"treats" templates only

• DrugMechDB-based. Also old notes here, here
• when Translator is ready to handle the data of TRAPI 1.4 validation error: gain / loss of function qualifiers for Gene -> Disease #699, we could add those templates back
• Our templates have Gene and Pheno as intermediates. Can we use other intermediates like Pathway or SequenceVariant (problem is lack of connections from those to Chem)?
• review: we cannot handle "opposite phenotypic features" right? An old ask is "a disease has phenotypes X; what drugs have the opposite of X as their side effect/adverse event" because we want drugs that can "counter" the phenotypes

"chem affects gene" templates only

• review interacts_with MetaEdges TRAPI 1.4 validation error when interacts_with predicate is in QEdges or Edges #701. I previously used interacts_with because of data in DGIdb and MyChem
• old ideas

Keep in mind

• UI has a 3-hop limit and subclass edges from node-expansion will count as 1. So templates should be only 1 / 2 hops right now...(first recorded here)