Using DocumentReferences, gather sample, assay, participant, and study information.

[RobertJCarroll]

For each assay result, we want to gather the following:

1. File ID to link with record
2. Tissue type from Sample
3. Condition / Disease linked to Tissue
4. Study name
5. Individual Sex

If the information is missing (eg, no sample links in KF), the data can be left as NA. We will post-process to generate as detailed of labels as possible.

[bwalsh]

# get the data
df_include =query_fhir_include(query_statement)
df_kf = query_fhir_kf(query_statement)
df_gtex = query_fhir_gtex(query_statement)

# dataframe looks like
#   columns: 
#       document_reference_attachment_uri (either drs:// or gs://), 
#       drs_uri, (if it exists)
#       document_reference_reference, ( DocumentReference/1234 )
#       file_path,  (downloaded document_reference_attachment_uri on local file system)
#       specimen_bodySite, 
#       condition_code,
#       research_study_reference, (full uri of research_study https:/example.com/fhir/ResearchStudy/1234)
#       patient_reference, (full uri  https:/example.com/fhir/Patient/123)
#       specimen_reference, (full uri  https:/example.com/fhir/Specimen/123)
#       ... extra columns (eg. observations) allowed
#   index:
#       document_reference_reference

pca_df =  df_include + df_kf + df_gtex

# go do PCA!

[mattions]

@RobertJCarroll plan ^

[torstees]

To get disease status associated with HTP in INCLUDE:

https://include-api-fhir-service.includedcc.org/Observation?_tag=HTP&code=MONDO:0008608

For Controls and

https://include-api-fhir-service.includedcc.org/Condition?code=MONDO:0008608

For Cases

That MONDO code is Down Syndrome. The other studies may be done differently, then Meen can clarify how those. His may all be Conditions with a different verificationStatus

[bwalsh]

gtext_v8 data frame here: gs://fc-be286b9f-3acf-4168-af6e-592df509391d/gtex_v8-dataframe.tsv