Rest API V2.0 - Testing feedback

[sajo-ebi]

Curators have provided their feedback on the rest API v2 testing . The link to the feedback document is below

• Some traits could not be seen in the payload for certain studies:
  GCST90243953, efoTraits is null, diseaseTrait not shown at all
  Note, this EFO (outer nuclear layer thickness measurement, EFO_0803371) also does not return any results in the efoTraits endpoint
  GCST003922: EFO was null - but later visible for same search

• Feature requests

Add cohort to payload
Add cohort to searchParams
Add gxe to searchParams
Where “platforms” is shown in the payload, snpCount and imputed are also displayed as separate fields, but not array manufacturer.
"platforms": "Affymetrix [360811]"
Array manufacturer could be accessible as a separate field (like snpCount and imputed)

1-3 are high priority, 4 can be left to see if users request it.

• SNPs
  Feature requests

Consider adding more info that is currently shown in UI: MAF, minor allele, etc. This info comes from Ensembl, so in theory the user can also get it from there, but for convenience we could include it, if it doesn’t add too much complexity to the design.

• Publications
  Comment
  We display a lot of granularity on authors (first name, last name, orcid, affiliation). Do we need to show all of this? Affiliation is not currently displayed in the UI. Could be more useful to show the sorted list of authors as displayed in the UI. Consider whether any of this is used internally.
  Potential external use-cases are:
  User wants to extract the full citation, including sorted list of authors
  User wants to search by first author
  User wants to search by any author

• efoTraits
  Bug
  Could not get response for any term using /v2/efo-traits/{efoTraitId} - unclear what is required for input here, Sajo thinks could be the internal ID, from the user perspective should be shortform.
  Worked ok for /v2/efo-traits/ searching by any term, shortform, trait etc

• Associations
  Question
  /v2/associations/{associationId} - requires internal association ID - is this useful for external users? (note, it is in current API). Discussed with Sajo: This needs to be a unique record, users can obtain it by first searching associations endpoint by some other parameter.

Other questions that came up from curators:

Do we support searching for associations below a certain p-value? (currently or in future) Could be a nice-to-have.

Does the v2 REST API support the use-case: I want to search for studies with a specific ancestry label (e.g. African), I want to search for studies with a specific sample size.
It looks like currently the ancestry/sample info is only returned in the long form (concatenated) and not in structured format.
In current REST API it appears in study endpoint like this:
"ancestries" : [ {
"type" : "initial",
"numberOfIndividuals" : 31135,
"ancestralGroups" : [ {
"ancestralGroup" : "European"
} ],
"countryOfOrigin" : [ ],
"countryOfRecruitment" : [ {
"majorArea" : "Europe",
"region" : "Northern Europe",
"countryName" : "U.K."
I think this is covered by Yomi’s comment, but please confirm:
Ancestries endpoint: This is a new endpoint, this is required to ensure scalability, it was previously embedded in the studies endpoint, but now a child object to the Studies data, and should still be accessible from within the study endpoint as a link on the URL

Rest API v2.0 feedback