Collaborative efforts produce clinical workflows for fast, translational genetic analysis

With individualized medicine—one of the holy grails of modern healthcare—diagnosis and treatment of patients would rely in part on each individual’s specific DNA profile, enabling truly personalized care. But in order for genetic information to contribute meaningfully to patient care, DNA testing has to be affordable and efficient. In 2017, the Mayo Clinic Center for Individualized Medicine (CIM) and the University of Illinois at Urbana-Champaign embarked on a two-year Grand Challenge under the auspices of the Mayo Clinic & Illinois Alliance for Technology-Based Healthcare with the goal of making DNA analysis a possibility for every patient. The first aim of the project focused on finding faster methods for clinical analysis of the whole human genome.

The Grand Challenge project, led by Eric Wieben, Ph.D. at Mayo Clinic and Matthew Hudson, Ph.D. at Illinois, tasked Liudmila Mainzer, Ph.D., Technical Program Manager of the Genomics group at Illinois’ National Center for Supercomputing Applications (NCSA), with speeding up clinical testing. Her group conducted analyses to find the fastest tools for genetic variant calling, which analyzes how a specific DNA sample differs from a standard reference. Ultimately, Mayo Clinic decided to adopt a new variant calling software that completes analysis 44 times faster than the traditional industry-standard pipeline—requiring just a few hours to process a whole genome, rather than days. But while faster software makes a significant difference, the bulk of the project lay in the next step for Mainzer’s team: wrapping the newly adopted software tools into a modularized clinical workflow. The resulting “Mayomics” (Mayo + genomics) variant calling workflow will be easy to maintain, update, customize, and run across Mayo Clinic’s many labs and numerous specialized procedures.

Nate Mattson, an IT lead analyst for the Department of IT Executive Administration at Mayo Clinic, coordinated with NCSA and with twelve clinical labs at Mayo Clinic to make sure the finished workflow would meet the needs of hundreds of clinical staff members. Mattson notes that the ability to configure and scale the workflow across multiple procedures and inputs, including whole genome data, is a critical design element—as is automation, which “enables 24/7 processing…without any human intervention” once samples have been sequenced. Modularity—separating tasks into self-contained scripts that can be mixed and matched as needed—is essential on multiple levels, Mainzer adds. “With so many different assays for so many different diseases and conditions, it would be impractical to write, test, and maintain individual workflows for each of them, and keep them in sync and up-to-date as the field evolves. Our design specifically addresses this through modules that can be used in hundreds of workflows, but only need to be updated once when changes occur.”

While the completed workflow satisfied the requirements set forth in the Grand Challenge, Mayo Clinic and Illinois decided to extend their collaborative project in order to add more functionality and configure new workflows, such as variant calling for tumor samples. In the meantime, the first Mayomics workflow has completed a process of rigorous testing by Mayo Clinic’s Software Quality Assurance (SQA) team and is now undergoing a “verification” phase in Mayo Clinic labs prior to official clinical deployment. Clinical work requires robust code and quality control, notes Mainzer, and has to meet exacting external specifications. According to Mattson, “Mayomics will support and exceed all of the auditing requirements set forth by CAP/CLIA and NYS/CLEP,” two sets of national standards for laboratory work.

Mattson and a team comprising Mayo Clinic research IT, clinical IT, and SQA staff work closely with NCSA Genomics to define and clarify requirements for each new request, confirm implementation details, and troubleshoot potential snags. The Mayo Clinic team also develops pieces of the workflow that are highly specific to Mayo Clinic internal systems and procedures. The collaboration works well: “There is value in clinical teams focusing on their own clinical process and medical informatics [while] someone else worries about code organization, workflow development and functionality,” says Mainzer. “These are two different mindsets and it helps when different heads are busy with each one.” Mattson is happy that the Mayomics workflows will enable more efficient, cost-effective analysis. “A diagnosis can be life-changing,” he notes. “Anything we can do to expedite that process without compromising quality is critical.”

This work was a product of the Mayo Clinic & Illinois Alliance for Technology-Based Healthcare. Major funding was provided by the Mayo Clinic Center for Individualized Medicine and the Todd and Karen Wanek Program for Hypoplastic Left Heart Syndrome. The Interdisciplinary Health Sciences Institute, Carl R. Woese Institute for Genomic Biology, and the National Center for Supercomputing Applications also provided support and resources.

Disclaimer: Due to changes in website systems, we've adjusted archived content to fit the present-day site and the articles will not appear in their original published format. Formatting, header information, photographs and other illustrations are not available in archived articles.