colleenXu
commented
1 year ago Discussion topics:
1: current BTE has buggy behavior when parsing the output of this API: see here, and here. I therefore suggest NOT "fully deploying" this as a KP used by BTE and others (by adding to BTE's config list) until this has been addressed
- It's difficult to write all the possible x-bte annotation because the bugs interfere with successful testing
- I anticipate that BTE could handle this better with the JQ-related features (two main issues) or a custom api-response-transform (current handling method). So I suggest getting the JQ-related features working first...
2: writing up all the possible x-bte annotation (depends on 1)
- ask the RaMP team: what is going on with the chemical/metabolite + gene namespaces? Ideally, we want to query for/retrieve as many entities as possible without getting "duplicates".
- right now, I'm assuming HMDB and entrez/NCBIGene are the only namespaces needed.
- Other chem namespaces returned in responses are CAS, chebi, kegg, pubchem.
- Other gene namespaces are ensembl, uniprot, hmdb
- an old note:
for example, these equivalent IDs for glutamic acid retrieve the same data: PUBCHEM:33032,CHEBI:16015,HMDB:HMDB0000148,KEGG:C00025
- right now, I'm assuming HMDB and entrez/NCBIGene are the only namespaces needed.
- Also ask RaMP team or search: what are the sources for the MetaEdges/operations? If the answer isn't simple or RaMP is the source, we can set RaMP as the primary source when we add it to the config list (noted here)
/pathways-from-analytes: based on old notes starting here- hmdbChem - reactPathway (already written)
- hmdbChem - wikiPathway
- hmdbChem - keggPathway
- entrezGene - reactPathway (already written)
- entrezGene - wikiPathway
- entrezGene - keggPathway
/analytes-from-pathways: based on Andrew's notes here- reactPathway - hmdbChem
- reactPathway - entrezGene
- wikiPathway - hmdbChem
- wikiPathway - entrezGene
/common-reaction-analytes: based on old notes here. I'm not clear on what this data is representing: provide gene -> chem (gene2met) and chem -> gene (met2gene) that are involved in the same reaction? what reaction/pathway they're both in isn't provided- hmdbChem - entrezGene
- entrezGene - hmdbChem
3: discussing with the RaMP team: what SmartAPI yaml file to have x-bte annotations on and register in the SmartAPI registry (including what repo).
- This probably involves removing the current SmartAPI registrations: older (alpha-version??) version, and my current (forked) version
- previously noted here
Opening an issue here to better track the status of this effort. In the future, when data resources reach the x-bte annotation part, we'll want to open an issue in this repo for tracking.
Previous discussion in https://github.com/biothings/pending.api/issues/69, with the currently-relevant comments starting https://github.com/biothings/pending.api/issues/69#issuecomment-1473339627
Related to post-processing and JQ-related work, which is described in https://github.com/biothings/pending.api/issues/69#issuecomment-1473375927 and https://github.com/biothings/biothings_explorer/issues/489#issuecomment-1480232931
I'm writing x-bte annotation in my fork: https://github.com/colleenXu/RaMP-Client/blob/x-bte-annotation/libs/features/ramp/ramp-api/src/assets/data/ramp_openapi_with_extensions.yml
That file is registered https://smart-api.info/registry?q=ac9c2ad11c5c442a1a1271223468ced1