false
zh-CN,zh-TW,en,fr,de,ja,ko,pt,es,th,vi
Catalog
Curating the Clinical Genome 2024 ePoster Gallery
Enhancing the AI-readiness of gnomAD v4 with Integ ...
Enhancing the AI-readiness of gnomAD v4 with Integrated GA4GH Genomic Knowledge Standards
Back to course
Pdf Summary
A computable model was created to exchange gnomAD data as evidence under the GA4GH standard, enhancing AI-readiness for gnomAD v4. Developed by a team from Broad Institute, Nationwide Children's Hospital, and Massachusetts General Hospital, the model focuses on the cohort allele frequency (CAF) to standardize data for variant interpretation workflows. The project aims to apply GA4GH genomic knowledge standards to gnomAD, the largest genomic evidence repository, and enable sharing of AI-ready genomic knowledge. The CAF model, designed as a nested structure, allows for common and extensible attributes across cohorts and subcohorts. The development involved implementing the VRS-Python VCF annotator at a gnomAD scale, selecting GA4GH VRS as a standard for unique variant identification. The model includes functions to extract GA4GH CAF Evidence and serialize digests for nested VRS objects, ensuring compatibility for downstream integration and reuse of curated data. This work represents a significant step towards facilitating the sharing of computable genomic knowledge within the genomics community. Additional details about the project can be accessed through a provided QR code.
Keywords
computable model
gnomAD data
GA4GH standard
AI-readiness
Broad Institute
Nationwide Children's Hospital
Massachusetts General Hospital
cohort allele frequency
VRS-Python VCF annotator
genomic knowledge standards
© 2025 American College of Medical Genetics and Genomics. All rights reserved.
×