The German Human Genome-Phenome Archive (GHGA) is funded by the German Research Foundation (DFG) within the framework of the National Research Data Infrastructure (NFDI).
The main goal of GHGA is to address the needs for establishing a national genome archive for human genome data while tackling challenges in data access, sharing, and privacy. GHGA aims to provide a national infrastructure for secure and safe storage,access management, dissemination, and analysis of human omics data under a coherent ethical-legal framework.
This repository aims to serve as the source of truth for the metadata schema used by the German Human Genome-Phenome Archive.
The schema is modeled using LinkML and exists as a YAML, which can be
found in src/schema/.
Using the YAML, and the LinkML framework, we autogenerate vendor/technology specific artifacts like:
- JSON Schema
- JSON Schema definitions of all classes and properties
- Python Dataclasses
- Python Dataclasses for each defined class
- Pydantic models
- Pydantic models for each defined class
- SQL DDL
- SQL Data Definition Language to create SQL tables for all classes
- SQLAlchemy classes
- SQLAlchemy classes to map SQL tables and columns to each defined class and slot, respectively
- GraphQL
- GraphQL types for each defined class
- ShEx
- Shape Expressions for each defined class
- RDF
- The model represented as RDF Turtle (TTL)
- OWL
- The model in Web Ontology Language (OWL), serialized as RDF Turtle (TTL)
- CSV
- A CSV summary of the model and its classes
Changes made
- has_publication removed from Project
- has_study made as list in Experiment
- introduced target regions as mandatory in Library preparation protocol
- Made library preparation kit retail name and manufacturer as optional
- Deleted has anatomical entity from Biospecimen
- Deleted has disease from Biospecimen
- Deleted has phenotypic feature from Biospecimen
- has_individual made mandatory under Biospecimen
- Introduced description to Analysis
- Introduced description as mandatory to Data Access Policy
- Introduced author, year and journal to Publication
- Defined slots for author, year, journal