BDC Variable Library
A LinkML-based variable library for BioData Catalyst datasets.
Background
BDC Variable Library (BDCVL) groups related variables into a single
LinkML slot per concept, so atrial fibrillation (MONDO:0004981) is one slot
with a comma-delimited source_study / dataset_accession /
variable_accession annotation, rather than separate near-duplicate records
per study. Concepts are grouped into classes that follow the BioData Catalyst
Harmonized Model (BDCHM), spanning all contributing BioData Catalyst studies.
Classes
Each class is a BDCHM class, with its own module, and its slots are the concepts derived as that class. Counts per class and per study are on the Stats page.
| Class | Concept identified by | Module |
|---|---|---|
Condition |
condition_concept |
bdc_varlib_condition.yaml |
MeasurementObservation |
observation_type |
bdc_varlib_measurement_observation.yaml |
Observation |
observation_type |
bdc_varlib_observation.yaml |
SdohObservation |
observation_type |
bdc_varlib_sdoh_observation.yaml |
DrugExposure |
drug_concept |
bdc_varlib_drug_exposure.yaml |
Procedure |
procedure_concept |
bdc_varlib_procedure.yaml |
Use the Schema tab to browse every class and slot.
From class to concept to variable
The library has three levels:
| Level | LinkML construct | Example |
|---|---|---|
| Class | class | Condition |
| Concept | slot, named after its ontology code | mondo_0004981 (afib) |
| Variable | annotation entries on a slot | one dbGaP phv / pht pair |
A concept's slot records every contributing study, and each study can have several dbGaP variables behind it (one per exam wave, hospital form, or follow-up survey). Those are stored as comma-delimited lists in the slot's annotations rather than as separate slots. For atrial fibrillation:
mondo_0004981:
title: afib
range: ConditionStatusEnum
annotations:
concept_id: MONDO:0004981
concept_name: afib
bdchm_class: Condition
source_study: ARIC, CHS, COPDGene, FHS, HCHS, JHS, MESA, WHI
dataset_accession: pht004037, pht004038, pht004039, ...
variable_accession: phv00203303, phv00203306, phv00203309, ...
source_file: ARIC-ingest/afib.yaml, CHS-ingest/afib.yaml, ...
exact_mappings:
- MONDO:0004981
One slot pools many dbGaP datasets and variables from several cohorts into a
single concept. The Schema tab lists the full
dataset_accession and variable_accession lists on each slot page.
Concept names
Slots are identified by ontology code (concept_id, also in
exact_mappings). The concept_name shown next to it is currently taken from
the source file name (for example afib, hist_cor_angio) and will be
replaced with ontology labels in a later release.
Structure
bdc_varlib.yaml: the umbrella schema; imports every class module.bdc_varlib_common.yaml: enums shared across classes (for exampleConditionStatusEnum), so every module is self-contained and can be loaded standalone, not only via the umbrella.bdc_varlib_<class>.yaml: one module per BDCHM class.
Provenance
Every slot's annotations record exactly which studies, dbGaP datasets
(pht), and dbGaP variables (phv) it was derived from, and the source file
it came from. The concept code (MONDO, OBA, OMOP, RxCUI, ATC, HP and others)
is carried through as a LinkML exact_mappings entry.
Source
Generated from RTIInternational/NHLBI-BDC-DMC-HV.