AstraZeneca’s Genomics Data Processing Solution Runs 51 Billion Tests in 1 Day on AWS

In addition to quickly gleaning insights from genomic data, AstraZeneca wanted to reallocate resources to scientific exploration and avoid spending bioinformatics time performing relatively low-value data management activities. Because the company collects petabytes of data through multiple sources in large bursts, it needed powerful, scalable compute capacity.

Having built on AWS before, AstraZeneca decided to expand its use of AWS tools to develop a cloud-based bioinformatics solution for rapid genomic processing and analytics.

AstraZeneca’s high throughput solution performs many steps of genomic data processing and analysis. Genomic secondary analysis looks at raw sequencing reads to reconstruct a genome and identify genomic variants that can be analyzed further in later stages. To automate the data orchestration of those steps, the architecture uses AWS Lambda, a serverless compute service that enables users to run code without provisioning or managing servers. To build out a task execution layer, the architecture uses AWS Batch, which dynamically provisions the optimal quantity and type of compute resources—such as CPU or memory-optimized instances—based on the volume and specific resource requirements of the batch jobs submitted. Along the way, the architecture sorts data into buckets using Amazon Simple Storage Service (Amazon S3), an object storage service.

On AWS, AstraZeneca sped up and improved productivity with scale, compute power, and access to rich technology services. Because the company can run analysis at scale when needed, data is available for analysis sooner. “We can now run over 51 billion statistical tests in under 24 hours, studying the effects of individual mutations or individual genes, each with a broad range of phenotypes,” says Slavé Petrovski, VP, head of Genome Analytics and Informatics at AstraZeneca’s Centre for Genomics Research. The company’s efforts are paying off. “We’ve provided genetics input into more than 40 of AstraZeneca’s drug discovery projects in 2020 using these capabilities,” says Petrovski.

AstraZeneca produced a rapid, efficient genomics bioinformatics pipeline that gives its scientists the time and resources to pursue innovation. As a result, the company’s Centre for Genomics Research is moving forward in its goal to analyze two million genomes by 2026.