The COVID-19 crisis demonstrates a critical requirement for rapid and efficient sharing of data to facilitate the global response to this and future pandemics. We can address this challenge by making viral genomic and patient phenomic data FAIR, and formalising it to permit seamless data integration for analysis. Phenopackets is a standard file format for sharing phenotypic information that facilitates communication within the research and clinical genomics communities. The OMOP model allows for large-scale analysis of distributed data to generate evidence for research that promotes better health decisions and better care. This gathered data is used by epidemiologists to monitor the infection, model it and make outbreak analysis and predictions to evaluate policy interventions. To harness machine-learning and AI approaches to discover meaningful patterns in epidemic outbreaks, we need to ensure that data are FAIR. To leverage data for federated learning/analytics, datasets can be discovered in FAIR Data Points; FAIR data repositories that publish human- and machine-readable metadata for data resources. This project aims to enhance interoperability between health and research data by mapping Phenopackets and OMOP and representing COVID-19 metadata using the FAIR principles to enable discovery, integration and analysis of genotypic and phenotypic data.
Covid-19 Data Platform Federated Human Data GA4GH partnership Interoperability Platform Machine learning
Project Number: 36
EasyChair Number: 63
Núria Queralt Rosinach ([email protected])
Phenopackets/OMOP mapping model. (4 days) Metadata extension of COVID-19 FAIR Data Points for federated Machine Learning. (1 day) Create a workflow to evaluate how mapping and metadata extension helps AI to discover interesting patterns. (2 days) Evaluate the mapping effort for semantic phenopackets developed in the EJP RD to OMOP and or HL7/FHIR-RDF.(1 day)
Phenopackets experts OMOP experts Clinical researchers Genomics researchers EGA experts GA4GH Beacon API experts Genotype-Phenotype biomedical informatics researchers AI/ML researchers
Number of expected hacking days: 4 days