Junior Hub Meeting - Conexión BCB (CSIC) - Oral presentation

Slides from my last Talk about Federated Genomic Data Discovery Infrastructure tittled "Deployment of a federated Beacon v2 instance integrated with an affiliated EGA node for genomic data discovery at ISCIII-CIBER"

Transcript

1. Deployment of a federated Beacon v2 instance integrated with an affiliated EGA node for genomic data discovery at ISCIII-CIBER Genomic Data Discovery for the Go-IMPaCT Cohort at ISCIII-CIBER Magdalena Matito Silvero²*, Alejandro Bernabéu Duréndez², Sara Monzón Fernández³, Isabel Cuesta de la Plaza³, Beatriz Pérez-Gómez¹², Pablo Fernández Navarro¹² ¹Centro Nacional de Epidemiología (ISCIII), ²CIBERESP, ³Bioinformatics Unit (BU-ISCIII)
2. I. Challenge: Genomic data discovery at scale Massive growth of genomic data The initiative • 18,000 participants from the Spanish IMPaCT cohort • Massive influx of population-based genomic data. System hosting-data requirements A federated architecture that allows external researches to securely query for data existence before requesting formal access to raw files At ISCIII-CIBER Researches must discover genomic variants across distributed networks Data needs to be discovered without centralization & compliance (GDPR and local governance)
3. II. Solution overview: a federated architecture Beacon v2 Instance Federated Discovery Affiliated FEGA node Secure and storage access Function Secure, long-term storage of sensitive datasets. Function Discovery of genomic variants in specific datasets through federated GA4GH API Data handled Raw encrypted files (FASTQ, CRAM, VCFs) + aggregated metadata. Data handled Aggregated VCFs Access Protocol Strictly controlled; requires formalized data access committee (DAC) approval Access Protocol Discover of data existence prior to access requests.
4. Affiliated FEGA Request Process Discovery: We have this dataset available (description) and you can’t see the metadata. Access requires approval from our Data Access Committee. External requesters also can apply for access. The Affiliated FEGA node, through its DAC, decides whether to grant or deny access. Data delivery is fully managed by the Affiliated FEGA node ISCIII-CIBER III. Affiliated FEGA Shared Data & Technical Considerations
5. Types of Metadata in FEGA Discovery Metadata Information describing the study as a whole: - title - description - research objectives - scientific context - keywords Legal Metadata Governance and access control information: Data Access Committee (DAC) and data access policies. Sample-level/Aggregated Metadata Technical and descriptive information specific to each element: - experiments, - samples, - datasets, - analysis - sequencing runs. Submitter Portal Required sample-level metadata: - Alias - Biological Sex - Subject ID - Phenotype III. Affiliated FEGA Shared Data & Technical Considerations
6. Data Sharing Models Aggregated Data (our case) Allele Frequency Counts model implemented in EGA • Show cohort aggregate information • Expose variant counts and allele frequencies • Generate derived cohorts from individual VCFs using Beacon-ri-tools Sample Level Data • Direct ingestion of individual VCFs • Requires stricter access control • Can be transformed to aggregated data via Beacon ri-tools 3 Datasets Access Levels Public Dataset accessible to general public Registered Restricted to approved registered users Controlled access Only registered users with additional node authorization IV. Beacon v2 Shared Data & Technical Considerations
7. IV. Beacon v2 Shared Data & Technical Considerations Result type selection Switching between Genomic Variants, Individuals, Biosamples, Analysis, Cohorts, Datasets and Runs 1 Filtering Terms Direct integration of extracted ontologies for complex phenotypic queries. 3 Response Summary Inmmediate, high-level dataset response hits before diving into controlled details 4 2 Genomic Query Builder Structured input allowing precise searches without requiring API coding Allele Frequency Beacon Browser
8. V. Strategic Alignment & Interoperability Local Node (ISCIII-CIBER) Deploying the affiliated Federated EGA (FEGA) node & Beacon instance for Go-IMPaCT data In collaboration with CRG Joint development ensuring alignment with ongoing EGA/FEGA ecosystem updates
9. VI. Results & Conclusions Full self-governance Sensitive data (FASTQ, CRAM, VCF) never leaves the ISCIII-CIBER storage. Complete GDPR and DAC compliance is strictly maintained. Rootless Security Podman-based containerization and Apache TLS reverse proxy to ensure robustness and isolated security from external attacks Standardized Interoperability GA4GH Beacon v2 and BFF standard to ensure the Go-IMPaCT cohorts align with the rest of Europe nodes Population-Scale Discovery 18,000 genomes become actively and safely discoverable to global researchers, accelerating precision medicine and rare disease identification.
10. Acknowledgements & Questions • This work was developed within the framework of the Go-IMPaCT Cohort Genome Project (PMP24/00005). • We acknowledge the collaboration of the Centre for Genomic Regulation (CRG, Barcelona) and the European Genome-phenome Archive (EGA) team. • We also thank team members of the BU-ISCIII including IPs: Alejandro Bernabéu Duréndez, Sara Monzón Fernández, Isabel Cuesta de la Plaza, Beatriz Pérez-Gómez, and Pablo Fernández Navarro. Contact Information • Magdalena Matito Silvero: mmatito@externos.isciii.es • Bioinformatics Unit (BU-ISCIII): bioinformatica@isciii.es