Supercharge your research with open science!

2 minute read

Written by: Lauren Wolfe -

The Coop team is excited to announce an upcoming series of posts that highlight ways that research labs can improve data management, analysis, and collaboration using open science tools and principles. These posts will walk through a ten key ideas from this Nature article. Each post will contain a motivating question, some longer term goals to keep in mind, readings, tasks, and discussion prompts. While this article was originally written as a multi-week plan of action, we encourage you to consider these ideas as a starting point for group-wide dialogue and actions surrounding lab culture, expectations, and facilitating open science.

Motivation

Analysis in scientific research tends to be solo work. Researchers, often self taught and feeling the pressures of publication, learn just enough to complete a few desired analyses. I know I’ve been guilty of creating freakishly long scripts with repetative copy-and-pasted code chunks and outputs scattered throughtout back early in my career as a bioinformatician. While these scripts might get the job done quickly, they are rarely reproducible long term. There is a growing need to establish good data science practices in scientific research, but it can be difficult to figure out where to start to develop sustainable and practical solutions.

The transition to more reproducible and open research requires a shift in perspective as well as a significant investment in team building and skill development, but the results are well worth it. If you’re uncertain about whether your group could benefit from more open science, try reproducing the results from data analysis you or your coworker did a year ago! This can be a great exercise to figure out areas for you and your lab to improve. Trevor Bedford has compiled some great information on the importance of reproducibility in computational work here. He also includes some recommendations for how to improve computational reproducibility in your lab group. If those recommendations seem daunting, don’t worry! This series of blog posts will walk you through the process as described in the Nature article.

Ten meetings to improve open science

Each post in this series corresponds to a topic for a group-wide discussion. These meetings should be encouraged by a team leader, but they don’t necessarily have to attend the meeting itself. Activities can span multiple weeks and should be revisited as needed, like when a new member joins the lab. The resources provided for each meeting are R focused but they remain illustrative of good theory and practice regardless of coding language.

You can view the whole series below (links will be added as the posts are published):

  1. What does a team workflow using open data science look like? What does the transition look like?
  2. How do we store and share our data, methods and code?
  3. What are our values? How do we build trust and facilitate collaboration?
  4. How do we streamline other aspects of our research, such as presentations and teaching?
  5. How do I collaborate easily with people who are remote or in different time zones?
  6. Which version of my code was I using, and where is it?
  7. I can’t understand or run other group members’ code.
  8. How do we record and organize data to streamline analyses?
  9. How do we help new group members to learn how we work, and how do we retain continuity when people leave?
  10. How do we continue learning and improving how we work together?

Keep an eye on the Coop Slack #general channel for updates on this series! You can expect a post or two per month and can always reference this post for links out to each post in the series!

You May Also Enjoy