The Open-Source Revolution in Life Sciences


Eric Nantz

Sr. Director - Innovation Statistics

Eli Lilly & Company

GVSU Alum!

@rpodcast

@rpodcast

eric-nantz

Discovery

GVSU (1998-2003)

  • STA215: I was hooked on statistics!
  • Balance of theory and practical use of quantitative methods
  • Minor in Computer Science
  • Statistical consulting capstone: immensely valuable

John Gabrosek & Eric Nantz (2003)

WMU (2003-2009)

  • Masters Degree in Statistics …
  • Convinced by important people to get a PhD!
  • Consulting & Administration for Stat Computation Lab
  • 2007: Internship at Eli Lilly (leading to my dissertation topic!)
  • 2009: Complete PhD program and joined Eli Lilly

Entering the Grid (of Life Sciences)

Clinical Statistician Role

Author statistical methods used in clinical trials:

  • Sample sizes required to meet desired treatment effect
  • Models & Inferences for primary and secondary outcomes
  • Gradually introduced to advanced methods (multiplicity control)

A New Challenge

Immunology Biomarkers

  • Proteomics, gene expression, DNA
  • Think differently:
    • Tailoring analytics to find optimal patient response
    • Use R for cutting-edge analyses
    • Launch computations on Lilly’s HPC cluster!

  • Ed, the lead from research IT

    Hello Eric, this is Ed from the research IT team. Did you submit a job array with ID 8675309?

  • Eric, the happy statistician

    Hello Ed, nice to meet you! Yes, I submitted that job earlier today. I guess it is taking longer than I anticipated to complete.

  • Ed, the lead from research IT

    Your HPC jobs have brought the entire HPC cluster down.

  • Eric, the statistician in complete shock

    Oh no!

A New Perspective

  • After correcting mistake, eager to learn better approaches to computing
  • Launched one of the most valuable collaborations in my career
  • Kickstart amazing innovations in our statistical computing infrastructure
  • Powered by open-source software!

Paving a New Path

  • Ever-growing R package ecosystem provided foundation for key statistical capabilities
  • Integration with HPC and web technology to move beyond static outputs
  • Taking (calculated) risks to challenge status-quo

Organic Collaboration

R/Pharma 2018-Harvard

Learning Out Loud

  • Launched the R-Podcast in 2012
    • Giving back to the R community in a novel way
    • Learning from the amazing package authors and practitioners
    • Conversations that profoundly impacted my journey

rstudio::conf 2019

Drivers to Success

  • Communication with cross-functional colleagues
  • Stepping outside comfort zone
  • Embracing innovation
  • Software Engineering best practices
  • Learning from others

Thank You!