Crowd-Sourced Science: The Open Source Wnt Project
This month, we spoke with Dr. Marian Waterman about her recently deposited 95-plasmid library of all 19 Wnt clones and how the library came to be. Dr. Waterman is currently the Professor and Vice Chair of Microbiology and Molecular Genetics at the University of California, Irvine (UCI) and the Associate Director of the UCI Cancer Research Institute. Along with her lab members and colleagues, Dr. Waterman started the Open Source Wnt Project as a way to pool community resources and efforts to develop an accessible and affordable set of reagents for the Wnt community. This crowd-sourced project is an example of how the nature of scientists’ collaborations has changed in an era of open source software and open access publishing. We spoke with Dr. Waterman about her choice to crowd-source the project, her lab’s experience with the project, and her advice to other scientists interested in crowd-sourcing their next experiment.
Q: For scientists unfamiliar with the Open Source Wnt Kit, can you briefly describe what the kit entails.
Sure, the kit is a complete collection of plasmid clones that contain full-length open reading frames (ORFs) for every human Wnt ligand. Wnts are secreted proteins that trigger Wnt signaling and there are 19 of these ligands in humans. We inserted all 19 ORFs into the Gateway plasmid system, which is a shuttle system that allows rapid and precise insertion of these ORFs into different plasmids for different purposes. The ORFs are full length, and their insertion is identical for every one of the Wnts such that expression of each from the plasmids should be identical. We also designed the kit so that each and every clone is readily excisable so that users are not restricted to the Gateway system. Every single clone has been verified at every step of cloning, and every ORF is present in multiple flavors in the Gateway system, bringing the size of the kit to a total of 95 plasmids. Some of the plasmid clones contain just the Wnt ORF, while other corresponding clones have the ORF-fused to an epitope tag. Since good antibodies for Wnts are hard to come by, the epitope tag allows one to follow each Wnt through modification and secretion inside the cell to its location in the extracellular space. The identical epitope tag also allows for quantitative comparisons between Wnts.
Q: What lead you to start the Open Source Wnt Project?
The first reason came from discoveries in our lab that different types of Wnt signaling affected the transcription factors that mediate signaling, LEF/TCFs, differently in the nucleus. We were trying to define the signal transduction pathway that was responsible. Wnts signal through two pathways: one is often referred to as “canonical”, or typical, and the other is often called “non-canonical”. Neither of those pathways fit with the signal we were defining, and that made it difficult during the manuscript review process – it was hard for reviewers to accept that there could be a blend of canonical and non-canonical signaling. During one of our lab meetings, we debated the reasons for why Wnt signaling is parsed so firmly into two separate pathways - a rigid classification that seems to have arisen over time. We wondered whether it is because everybody uses the same two Wnts to do their experiments (canonical Wnt3a and non-canonical Wnt5a). Maybe if everyone had access to all 19 Wnts, they could express the appropriate ligands that match the right receptors in their experimental systems. We wondered whether access to the right Wnt ligands would modify or even erase rigid classification of Wnt signaling.
The second reason for the project was practical. We wanted to clone the appropriate Wnts, but I didn’t have the money to do it right – to be complete and to do it well. So I approached my program officer at NIH and asked for supplemental support to do this very practical straightforward cloning project. While my program officer was very positive and gave me lots of moral support, there were no supplemental funds available. It was then either left up to me to come up with a creative way to fund this, or else to let it drop.
The third rationale was that I thought it would be interesting to do a social experiment. If people who want to study Wnt signaling, have easy, affordable access to the right reagents, would this expand the number of people looking at Wnt signaling? Would this change our view of Wnt signal transduction and would it change the rigid classification? To do the experiment you have to make the right reagents available.
Q: What was the reaction you got from other scientists when you began looking for people to participate in the project?
Reactions were immediate and extremely positive. That was very gratifying. Just like the experience with my program officer, when I approached people and described this project, there were only positive reactions. There was only one person who questioned why I would want to do this because it was a distraction from my regular line of work. One colleague who was extremely positive was Dr. David Virshup (Director Cancer and Stem Cell Biology Program, Duke University-Singapore). David offered to share the workload. So, although the open source idea came from my lab, he was an early supporter and devoted people and reagents to the project.
Q: What did you learn from your crowd-sourced science experiment?
So that’s the most fun part, actually. I like the Wnt scientific community; I think the sense of community has been kept healthy and dynamic by regular Wnt meetings in the U.S. and Europe and I thought I could tap into that gestalt. I found that when I called Wnt scientists on the phone and said “Would you help me financially with this project or would you send me your human Wnt clone?”- they responded positively. I then knew that we were doing something that people would value. I learned that crowd sourcing is like taking a vote or doing a market survey. You can quickly learn whether your idea is, or is not compelling.
The second thing I learned is that the money and resources I crowd-sourced was not nearly the amount that I needed to complete the project – not even close. It did help start the project, and that was good, and it helped me be confident that I should go forward because people were interested in the project. But it is probably always true that things cost much more than you think.
The third and final thing I learned is that crowd-sourcing makes people part of the project – it creates a little bit of ownership for everyone and that stimulates pay-it-forward behavior. For example, after we finished the Open Source Wnt Kit, we sent the collection of plasmids (at that time it was 76 clones), to everyone that had participated in the initial crowd-sourcing (that was the pay-off; everyone who helped support the project, received the entire collection of clones when the project was completed). In that collection, we discovered an issue with the epitope tag version of the Wnts. Epitope tagging inactivated the Wnts. While these clones were still useful (as silent markers of Wnt secretion), it was obviously not optimal. Dr. Xi He (Professor, Boston Children’s Hospital, Harvard Medical School) and Dr. Bryan MacDonald, recipients of the Open Source Kit, figured out why epitope tagging inactivated the Wnts. It was a problem with the amino acid sequence in the linker region between the ORF and the epitope tag. They corrected the problem in all 19 clones using site-directed mutagenesis and added the corrected clones to the Open Source collection. I consider that to be paying-it-forward - giving back to the community. And I also consider it to be in the spirit of “Open Source” which is to help make an openly available product better.
Q: What advice do you have for other scientists who are considering utilizing crowd-sourcing to aid their research?
I have 7 things: (1) When you are thinking of crowd-sourcing a project, you first want to assess whether it has broad appeal. If the project is something that only a couple of labs will be interested in, it will be hard to crowd-source to any significant level. Cloning all 19 human Wnts and designing the kit for multiple utilities and applications, has broad appeal.
(2) You want to be able to tell the person you’re approaching that they will get something in return. In this respect, crowd-sourcing is a little like Kickstarter – the for-profit company that hosts crowd-funding efforts. I suppose you could crowd-source without the promise of something in return, but it’s very appealing to people if they know six months or a year down the line they are going to get something back.
(3) Have a quality control plan. In our situation, we established stringent quality controls and this is the reason it cost so much. For every single cloning step, the ORF and flanking sequences were sequenced i both strands. We did this so that we never had mutations carried through generations of clones and plasmids. We still had problems, i.e. tags that inactivated the Wnts, but that was a design issue, not a mistake or a mutation. Crowd-source participants knew what our quality control plan was, and this demonstrated that we were serious about the project and that the reagents were trustworthy.
(4) It’s always more expensive then you think. You may think that raising “X” amount of money is sufficient. But if possible it’s better to raise a little extra.
(5) If you have a community where interested parties cannot help out financially, think about inviting them to help you - make it a group effort. My conversations with David Virshup evolved into a joint effort. Our groups had monthly Skype sessions where on his side of the screen were all of his lab members involved with the project, and on my side of the screen were all my members involved in the project. It was really fun to do the project this way and we became closer collaborators.
(6) Designing a workflow that everyone adheres to is extremely helpful when you have a community crowd-sourced project. This is especially true if there are a lot of routine tasks involved. A workflow plan helps keep everyone synchronized and makes troubleshooting a lot more efficient.
(7) Finally, if you’re going to crowd-source the making of a reagent or a tool, it’s a good idea to think about how you can put out a paper describing this reagent. A manuscript that can be referenced, like a methods or a tools paper is very useful. It was an extremely educational experience for my two graduate students, Rani Najdi, the first author of the paper, and Stephanie Sprowl. They were in charge of the Wnt team in my laboratory and they spent a good deal of time debating and designing the experiments that best demonstrate the utility of the kit. They then worked with Kyle Proffitt and others in the Virshup laboratory to carry out these experiments.
Q: How has “open source” and “open access” science (e.g. ResearchGate, PLOS) and social networking changed how your lab gets work done? How have these affected your ability to initiate collaborations or changed the types of collaborations that you now participate in?
The term open access is highly prominent in my laboratory. All of my students want to publish open access papers. None of them want to go to a journal that they know will cost $5,000-$6,000 to put out and that they won’t be able to retrieve once they leave the lab and don’t have access to a subscription. I think the models show that open access works and that it’s better for everybody. I would include your organization, Addgene, in that sentiment. Addgene is part of that inspiration, a non-profit that is trying as best it can to share research tools and reagents. Open source is a term that came from the computer science end of things and Linux is one of the best examples. My husband is a bioinformaticist and ardent advocate of Linux, open source software, and the EFF (Electronic Frontier Foundation). I took the “open source” term and applied it to the Wnt project because that was the intent. We want the Wnt plasmids to be easily available to everybody, and we hope that others will improve or expand the kit. We figured that just like open access publishing, or just like open-source Linux, this resource would be better in the long run. My opinion is that quality discoveries come from access to good tools and reagents. The wider the spread of these tools and reagents, the better the science. And something like Addgene and something like open-access publishing enables that to happen. This project hasn’t really changed my collaborations, except of course with the Virshup laboratory, but the idea of open-source and the idea of open access has permeated everything in the lab. We have Dropbox, that’s how we share data. We have journal clubs through Mendeley, which used to be the freeware referencing manager. And we are adopting open source software for data and image analyses.
Q: What is your lab working on now? And are there any new research questions or directions in the Wnt community that you’re excited about?
Well actually, my lab is not known for studying Wnt ligands. We just wanted to do the project so that we could get back to figuring out how our transcription factors are controlled by Wnt signaling. That is where we’re back to actually. We’re back to working on the LEF/TCF transcription factors, their role in cancer, and the signals that control their expression and localization. I don’t expect that I’ll be a Wnt ligand researcher. However what is exciting is that the first structure of a Wnt ligand binding to a receptor has been solved this last year. There are an increasing number of studies on the structures of the receptor on the cell surface. So for Wnt ligands, I think we are on the cusp of learning a lot about their structure. I hope that having this set of Wnt tools will contribute to these new developments. And while I expect it will take a couple of years, I want to watch how this collection of clones will be used. What will be the outcome of the social experiment? Will this collection of Wnts be useful to people and will it change what we know about Wnt signaling? I’ll also be looking out for whether this inspires anybody else to pay it forward.
Open Source Wnt Kit described in:
Najdi R, Proffitt K, Sprowl S, Kaur S, Yu J, Covey TM, Virshup DM, Waterman ML. A uniform human Wnt expression library reveals a shared secretory pathway and unique signaling activities. Differentiation . 2012. Sep; 84(2):203-13.