CSIRO builds serverless architecture to query 100x more genomic variants on AWS
Australia's CSIRO has developed Serverless Beacon (sBeacon), a serverless architecture that queries genomic variant data without copying large datasets into a central database. Built on Amazon S3, Lambda, Athena and DynamoDB, it implements the GA4GH Beacon standard and answers a 10,000-base-pair query in about 1.52 seconds at an estimated cost of $0.00013.
- VCF files stay in Amazon S3 while metadata is indexed separately via Athena and DynamoDB
- A 10,000-base-pair query averages 1.52 seconds at an estimated cost of $0.00013
- A 2,504-sample chromosome 1 dataset was onboarded in 18 seconds
- The design targets up to 40 million individuals and 3 billion genomic locations
Read next
AI