Harnessing St. Jude Computing Power at Home
St. Jude optimizes science with the power of cloud, internet resources.
By Chris Pennington
Computational biologists and other bioinformaticians use math and computer science to understand the building blocks of life and other biological systems. It takes immense computing power to process hundreds or thousands of genomes.
On the St. Jude Children’s Research Hospital campus, operating these systems is no problem. But when the institution shifted to a work-from-home model in the wake of the COVID-19 pandemic, scientists had to figure out how to move that same high-tech capability to low-tech kitchen tables and home offices. Traditionally, the computing power came from a high-performance computing cluster, or HPC. For several decades, an HPC cluster has been the model that worked well to process millions of computational tasks at St. Jude.
During the pandemic, many labs focused on computational work. At St. Jude, two systems are used to access and program all this data. Off-site investigators log in through a virtual private network (VPN) to access the HPC. The other option is a cloud-computing infrastructure through St. Jude Cloud.
“The St. Jude Cloud team is providing an option to our research community that is separated from their on-campus data and computational tools,” says Chief Information Officer Keith Perry. “We want to drive science to the cloud as an option for developing and furthering their work.”
Vast resources, now available via cloud
Clay McLeod, director of product development and engineering in Computational Biology, sees the cloud as an option for the new computational demand.
“Over the past several years, my team has primarily focused on sharing St. Jude data with the rest of the world,” McLeod says. “Given the recent disruption, we’ve turned our attention to helping our own researchers leverage vast computing resources available in the cloud—including near-infinite computing and storage capacity.”
So, McLeod and his Computational Biology colleagues worked with Information Services to help researchers try working in the cloud.
A step further
Perry says moving some of the computational and informatics resources to the cloud takes advantage of data that has already moved.
The St. Jude Cloud team has been hard at work making this happen.
“The founding concept of the cloud is that no one should need to fight over resources,” says Kirby Birch, senior product owner on the St. Jude Cloud project. “We wanted to see how we can take the cloud a step further and provide this resource and make it readily available.”
Birch’s team created specific introductory pathways for new users. Support was the first step to help new users navigate St. Jude Cloud. That meant developing an online channel for discussion and helping users overcome the uncertainty of using a new interface.
“The cloud team aims to make the transition as seamless as possible,” Perry says. “It could be the interfaces or simply moving to a completely new setup, but researchers want to be close to their data and the hardware.”
Supporting state-of-the-art research
“We’ve put our focus on helping the labs,” Birch says. “Our team has dedicated its time and effort to help researchers get started.”
At first, typical cloud users were those within Computational Biology, but Birch says they soon reached a point where they could help others across the institution.
“We’re available to work with labs across St. Jude,” she says.
St. Jude Cloud is not the best choice for every lab, but it’s a viable alternative for many that use large amounts of data or perform computational analyses at scale.
“Ultimately this is a collaboration between our team and Information Services,” says McLeod. “If a lab is interested in using the cloud, we’re here to help them get started. Then, all of our digital solutions work together to serve everyone efficiently.”