Translational Data Science

You can connect directly with our data science and software teams through our Data House Calls program.

The Translational Data Science group collaborates with and supports both research and clinical staff to integrate effective data science and analysis into research and clinical care. Our goal is to create a feedback loop to foster a learning healthcare system.

We develop data products, data packages and data analysis tools to enable effective data use and integration into research and clinical care. We also collaborate with our clinical and research partners to develop statistical/ML/NLP models, data analyses, and data-driven dashboards and visualizations using real-world healthcare and laboratory-generated data.

Clinical Data Science

Our team supports clinical data science at Fred Hutch by analyzing real-world data generated over the course of health care delivery, and by improving the interoperability, quality, and accessibility of these data. They collaborate with Hutch researchers and staff to understand critical questions about cancer patients, develop impactful and practical data analyses, and to develop tools that support democratized data access, exploration, and analysis.

Research Informatics

Our team supports the use of biomedical datasets in the research context that often include large scale public or private genomic data, licensed or regulated datasets (such as those under DUAs or legal protection like GDPR), or laboratory-generated data that require significant computational processing as part of their analysis. We focus on providing open-source tooling, best practices for computational workloads and data management, and other resources such as computational workflows and templates that staff can use as a jumping off point to customize for their own work. We work closely with Fred Hutch IT’s Scientific Computing group who provide support for our on-prem computing cluster, scientific data storage, and research applications.

WILDS WDL Development Program

We offer free, collaborative support to help Fred Hutch researchers develop WDL (Workflow Description Language) workflows for scalable, reproducible computational analyses. In exchange, completed workflows are contributed to the open-source WILDS WDL Library (GitHub) for the broader research community.

Through this program, we:

  • Collaborate with researchers to convert existing scripts, pipelines, or tool combinations into WDL workflows
  • Provide workflows optimized for Fred Hutch HPC and/or cloud environments
  • Publish all workflows to the WILDS WDL Library under open-source licenses
  • Provide documentation on how to run and adapt workflows

What to Expect:

  1. Submit a request using our intake form
  2. Initial meeting – We’ll schedule a Data House Call to discuss your project, assess fit, and scope the work
  3. Development – The WILDS team writes the WDL workflow
  4. Testing & feedback – You test the workflow with your data and provide feedback
  5. Publication – The workflow is added to the WILDS WDL Library with appropriate attribution

This program works best when you have existing scripts or a pipeline you want to scale up, your workflow has broad applicability to other researchers, you’re willing to have the final product shared publicly, and you can commit time to testing and providing feedback.

Ready to get started? Submit a request here.

Questions? Reach out to the WILDS team at wilds@fredhutch.org or on the #workflow-managers channel of FH-Data Slack.

Research Data Applications

We support the private Fred Hutch instance of cBioPortal in collaboration with Fred Hutch IT’s Scientific Computing group. cBioPortal allows users to easily explore multidimensional cancer genomics data with an intuitive point-and-click interface. For more information, read the SciWiki documentation and visit cbioportal.fredhutch.org to try it out.