How I Work Open: Steven Roberts

Steven Roberts photo

Steven Roberts
Associate Professor, Aquatic and Fishery Sciences
Kenneth K. Chew Endowed Professorship in Aquaculture

Spotlight on:
Open Science

What kind of open work do you do?
We try to do everything openly. The [open] lab notebooks are the big things so people can see the research as it progresses. We try to make preprints available, and we share via social media. We also record and publish lab meetings online.

Please describe your lab and its work.
I’m in the School of Aquatic and Fishery Sciences within the College of the Environment. In the Roberts Lab, we use genomics techniques to understand how the environment influences shellfish, so we do large DNA analysis and gene sequencing.

How would you describe your research to your family?
It’s become simply marine biology. I don’t want to go into too much detail because most people can understand that (marine biology).

What are the methods and tools you are using to make your work open?
Well, we use the internet! We publish everything so we can find it easily. We rely heavily on GitHub for lab communication, lab notebooks, and data analysis. We also use Jupyter notebooks to publish and collaborate with each other.

Another thing we’re getting ready to do …is an audio podcast “behind the paper,” to talk to students and authors about the process. One student is really excited about starting that.

Our grant proposals are open as well. Sometimes I forget what all we’ve made public. It’s been a standard way of operating for so long.

Do you provide instruction about how to use GitHub?
We are currently redeveloping our GitHub presence to create repositories that explain how to use GitHub. I teach some of my graduate-level classes how to use it, but some is just learn as you go.

What barriers have you faced in trying to work open?
I’ve thought about it a little but I can’t think of a barrier. It’s just easier to do everything openly. It’s less work.

How do you incorporate open methods into your teaching?
I’ve brought openness into the classroom, for example with open notebooks.  With benchwork classes and computational work they bring it into their own labs if they’re graduate or undergraduate students.

Are students receptive to working open?
Everybody seems to be super excited about it. In some of the classrooms, they might have to discuss it with their collaborators. With large projects it’s a matter of making sure everybody knows what’s going on. Sometimes with large projects used for classes we come to some sort of compromise on how much to share.

Why do you work open?
Honestly, working openly makes it easier for me. If I need to know how to do something I can go to google with a question and find it’s in our lab notebooks. It isn’t lost and is backed up. It facilitates collaboration: if a colleague has a question about my work, I can share the link to the notebook and it saves time.

There’s also the aspect that I’m paid by the state. More and more money comes from the federal government, and there’s a desire that members of the general public see what we do. They may not understand it all but we strive to make it available and parts of it understandable.

Others say it’s for my “future self,” too. If I’m working on a paper that’s two or three years old today, backing up is good but there’s extra value in making things open so we can get back to it easily.

How did you get started working open?
I realized there was no barrier to using the tools so why not use them? When I was in graduate school I’d create these websites where we’d make data and biological reagents—which wasn’t even data then—available. Ever since then the research has changed. We deal with a lot of data. We know that others have far more interesting questions working with these datasets that we won’t even be able to tackle with this data. We don’t want graduate students spending resources and time to reinvent the wheel when they can make greater advanced research discoveries based on what we’ve done.

When you were in grad school, was working openly encouraged or is this a new thing?
There wasn’t even Google back then. It was benchwork. The type of research we did was not that amenable but my advisor was enthusiastic about the websites we were creating. A lot of people have done original stuff in open science outside of fields you’d consider more traditionally open.

What opportunities does working open offer that traditional scholarship does not?
The main thing is you can build on the work. People can take an open dataset and build on [it]. People can do interesting things with it without your participation, which is great. On the other hand, prospective graduate students and other researchers are more likely to contact me and ask me questions which sets up nice collaborations. It reduces barriers for communication when people know everything is out there. We get feedback—I’m thinking about preprint articles—that makes our work better.

There is shift in thinking about research as a product to thinking of it as a process. We’re documenting it along the way. How does promotion and tenure catch up?
I think it is. The National Science Foundation has made changes to classify code and data as product. Tenure will catch up with this.

Most of your work is in teams. Is it problematic going through the promotion and tenure process to distinguish what your piece is? Is that a barrier?
You could argue that if work is open then everybody sees what everybody is doing. I collaborate with a lot of people who do open science. It benefits student collaborators. That’s the reason you do things openly: it’s very clear what your part is and what you’re doing.

Do you feel supported working open?
By the College of the Environment, certainly. They’re interested in outreach and education for people outside of the ivory tower.

What’s your vision for an open future?
Maybe in the future we won’t be talking about open science in interviews like this–it would just be how everybody operates.

The UW Libraries are working toward this by improving our data repositories. I’m excited about efforts toward an University-wide open access resolution at UW. I hope that goes forward.

There are probably a lot more ways the university could support open scholarship. The amount of data collected in all fields is going to get bigger. There’s plenty of opportunity for the University to assist. There’s the repository but there are a lot of other places where there could be support for this.

What do you think is the future of open scholarship in your field?
Openness could change our methods. Even if you don’t share data, sharing methods and how you do things helps everybody. I had a reviewer one time who said, “It’s not necessary to see how the sausage is made.” But I think it is. It helps younger scientists see that perfect papers don’t just magically appear. Maybe 80% of what we do doesn’t work and we should say that. Then graduate students would see that and not try to reproduce it. They should know about the “dirty work.” I don’t see a downside to it.  

What’s the vision we should be working toward as a community?
If you want to get started a simple thing — especially if you’re a student –is to document your work on a daily basis on a simple WordPress blog, which the University provides. There’s a low- to no-overhead way to document what you’re doing that way. While GitHub was made for software, it’s also great for writing, open science, and collaboration. Jupyter Notebooks is similar to GitHub and works a data analysis tool, [too].


Leave a Reply

Your email address will not be published. Required fields are marked *