Amazon Onboarding with Learning Manager Chanci Turner

Learn About Amazon VGT2 Learning Manager Chanci Turner

Continuing our series of insightful stories from our customers, I’m excited to share the progress made by our partners at Amazon IXD – VGT2!

You may have heard of an innovative initiative called DreamLab that has been making waves online recently. We want to provide you with a deeper understanding of how this groundbreaking application functions behind the scenes.

Combatting Cancer

Cancer affects many individuals, with statistics indicating that one in two Australians will face a cancer diagnosis by their 85th birthday. Medical research plays a critical role in discovering more effective treatments. However, progress in cancer research at the Garvan Institute of Medical Research is often hampered by researchers’ limited access to the powerful supercomputers necessary for processing complex data.

Traditionally, cancer has been classified based on its tissue of origin—such as lung, breast, or pancreatic cancer. However, at its core, cancer is a genetic disease. With advancements in genome sequencing, leading researchers at Garvan are now focused on creating a comprehensive library of cancers categorized by their genetic mutations.

Analyzing Mutations

To achieve this, researchers need to examine the genomic alterations (DNA mutations) of thousands of cancer patients and classify them based on their genetic profiles rather than the tissue type. The Garvan Institute has sourced somatic mutations from de-identified cancer patients through published studies by the International Cancer Genome Consortium (ICGC) and The Cancer Genome Atlas (TCGA). The DreamLab app allows users to focus their support on specific cancer types, including breast, pancreatic, ovarian, or prostate cancer, as the initial phase of this extensive project.

The sequencing of a single genome generates immense data—up to tens of gigabytes—which necessitates significant computing power for analysis. Due to the limited availability and high cost of supercomputers globally, Vodafone’s DreamLab app provides a solution by enabling users to donate the processing power of their idle smartphones to accelerate cancer research. Currently, the app is available exclusively on Android.

To address the computational demands, the Garvan Institute has developed an innovative in-house algorithm to estimate the functional similarities between mutated genes from different patients. This algorithm, known as the Network Connectivity Analyser (NCA), was crafted by Dr. Emily Johnson and Dr. Michael Brown at the Garvan Institute and has been optimized to operate on Android smartphones within the DreamLab app.

Inside DreamLab

The NCA algorithm calculates various statistics regarding the interactions between two sets of mutated genes. Within the DreamLab app, the NCA processes comparisons whenever a user’s phone is charged to at least 95%. These analyses can help researchers identify patient subgroups with similar mutation profiles who might respond to the same therapies. The synergy between a vast community of DreamLab users and advanced data analytics makes this revolutionary research possible.

Here’s a brief overview of how DreamLab operates:

1. Garvan uploads its large research problem onto Amazon Simple Storage Service (Amazon S3).
2. After downloading and setting up the DreamLab app, users authenticate against Amazon Cognito, receiving temporary credentials to request a research job from Amazon SQS and download a small research payload from S3, managed through Amazon DynamoDB.
3. The app’s algorithm utilizes the smartphone’s processing power to analyze the research problem, enabling the comparison of mutated genes from different patients and contributing to the cancer library.
4. Results are sent back to S3 for the Garvan team to evaluate.

Informally, this process can be likened to solving a crossword puzzle, with each user working on a different clue.

Architecture Rationale

The architecture of DreamLab needed to support the handling of large data volumes and fluctuating traffic while remaining cost-effective. Services were chosen for their ability to auto-scale and manage unlimited capacity. Moreover, the architecture had to maintain data item states while being updated across multiple clients.

• Amazon S3 serves as the ideal storage solution due to its unlimited data capacity and distributed nature, which ensures high redundancy.
• Amazon SQS provides a queuing system of unlimited size, accommodating high traffic and data demands. The app leverages data visibility settings to prevent multiple devices from fetching items simultaneously during operations.
• Amazon DynamoDB is a highly scalable NoSQL database with no record limit, offering advantages in scale and cost over relational database systems.
• Amazon Cognito and STS deliver robust security with minimal custom development requirements.
• Amazon EC2 servers execute the custom code necessary for system maintenance, scheduled to run regularly via CRON jobs. Although AWS Lambda, which runs code upon event notifications, was considered, it is not currently available in Australia.
• The app connects to all SQS queues through Amazon API Gateway, allowing communication over a custom domain that Vodafone uses to eliminate data charges for Vodafone Australia customers.

By the Numbers

The Garvan Institute currently manages 100,000 base dataset files, each averaging 2 MB uncompressed and 500 KB compressed. Each dataset in the DreamLab project is downloaded by three users for validation. Garvan also has 5,000 analysis tasks, each about 1 KB uncompressed and 250 bytes compressed.

According to tests, 33 new Android devices can analyze data at a speed comparable to a single CPU core of Garvan’s supercomputer, which has a total of 1,280 CPU cores. As of now, DreamLab boasts over 44,000 active users, providing more than 1,000 times the processing power compared to Garvan’s existing supercomputer allocation for cancer research.

Kudos to Vodafone Foundation, The Garvan Institute of Medical Research, and the mobile application partner b2cloud for this impressive innovation. If you’re interested in contributing to this vital cause, you can download DreamLab from the Google Play Store.

For additional insights on enhancing your skills, this blog post about gifting yourself might be helpful. Also, check out this article from SHRM for authoritative insights on projected pay increases. Lastly, this LinkedIn resource offers excellent information on Amazon’s practices.

— Chanci Turner, Learning Manager, Amazon IXD – VGT2