Julien Colomb on open science and firing lasers at flies

Continuing our ‘open science stars’ series, we’re happy to present Dr. Julien Colomb this week! Julien is a postdoc in Berlin, and we’ve been working together (well, Julien has tolerated my presence..) at Open Science meetups here, which he’s been using to build an active community over the last 10 months or so. He recently published a cool paper in PeerJ and built a new ScienceOpen Collection, so we asked for his thoughts and experience with Open Science!

Hi Julien! Thanks for joining us at the ScienceOpen blog. Could you start off by letting us know a bit about your background?

Hi John. My pleasure to be here. [We’ve known each other for a year and he still can’t spell my name..]

I have been interested in neurobiology since my high school time; I got to work with Drosophila during my Master’s thesis and could then not leave the field. I worked about 10 years on the neuroanatomy and behaviour in the fruit fly larvae and flies in Switzerland, Paris and Berlin. In 2013, I decided to stay in Berlin when the mentor of my second post-doc, Prof. Brembs, moved to Regensburg. In the last 3 years, I have been jumping between different jobs in Prof. Winter groups, I have been wandering in the startup community in Berlin (founding Drososhare GmbH), and trying to foster open science and open data. At the moment, I work half time at the Charite animal outcome core facility, while we work on getting a beta version of the Drososhare product (a platform to share transgenic Drosophila between scientists). I also run the Berlin Open Science Meetup.

When did you first become interested in Open Access? What was your initial reaction when you heard about it?

As a scientist, being in big institutions, I rarely had problem accessing any paper. When I heard about PLOS One (probably in 2010), I was interested in the idea of one single megajournal, and did not get what the open access movement was. It is only in the last two years that I got to understand what Open Access means, and now, the problems of library budgets I heard as a PhD student in 2005 became relevant to me! I am glad the Open Access war is over (even Elsevier state that they are pro-open access), now starts the war for a fair price for Open Access, and this will be a tough fight.

You recently had a paper published in PeerJ about motor learning in fruit flies. Can you tell us a bit about this and your ongoing research?

Brains are not like computer input-output systems, but work mainly as output-input systems (they first produce behaviour and then modulate the output depending on the input they get). In the lab, we aim at understanding how this system learns, especially why learning-by-doing is more effective than passive learning. Despite a century of research this is still puzzling. We therefore started at one corner of the puzzle and tackle the question of the biological bases (which neurons, which genes) of “pure operant conditioning”, something related to motor learning and where the brain works as a pure output-input system. In practice, we punished a fly (with a laser beam!) when she is trying to fly to the right, but left turning attempts are safe (the fly is not moving at all). After 8 minutes of training, flies are doing the safe behaviour more often than the previously punished one in a 2 min. test. In this paper, we have shown that this learning is dependent on the action of a protein kinase C in the motor-neurons of the fly. It is very interesting to see that neurons so close to the action performance are plastic, and that what we do do not depend only on our brain.

On the technical side, I had to build together a machine which can measure the torque force (yawn) the fruit fly is producing while flying (the fly is glued to a hook and cannot move). I also develop new algorithms to collect, archive, analyse and present the data in a couple of button pushes, mainly using R code.

Why did you choose this journal to publish your work in?

We wanted an Open Access journal, we had mitigated experience with PLOS One, good work with F1000 Research and wanted to test another one. Since I was writing the manuscript on Overleaf, we wanted to try PeerJ. We also got a membership account, such that we will be allowed to publish one paper a year free in PeerJ from now on. Since I may leave academia, I also thought it would be a way for me to publish other papers in the next years without having any funding.

How important is Open Access publishing to you at this stage of your career?

I made a statement in January. I think the ‘publish or perish’ culture and the impact factor games are rubbish and I will not play that game (even if it means leaving academia). Open Access is a must, and I am pushing it also when I am not first author. My “career” is getting out of academia, and for me the only thing which matters now, is do the right thing for science, even if this is bringing much for my scientific career. On the other hand, some people proved that going the open access way can be productive for your career.

I think the ‘publish or perish’ culture and the impact factor games are rubbish and I will not play that game

You also published the ‘negative’ results from your research. Why did you do this?

The good question is why would I not do it. I worked hard to obtain these negative results, there is information we can take out of these experiments. It is not because I failed to find the answer to the question I had that my attempts have no value.

Why do you think many researchers choose to with-hold publishing their ‘negative’ results? What does this say about the current system of scholarly publishing?

In most cases, it is just not worth the work and that is as simple as that. On one hand, negative results cannot be published in high standard journals, on the other hand, the time to invest to publish them is a least as high as for positive results. In my case, the cost of adding the negative results was not significant (the data was analysed and the graph were produced using the same code as for the positive results), and the benefit was … well we will see if there is any.

You also made all the data and code supporting your research available via Figshare (details here). Why did you choose to do this? How important is it to you that all outputs of research be made available and re-usable?

I took that question very seriously when I started to collect data. We wanted to do our science more open, but were lacking the tools to do it. At the end of our journey, I had learned to use R and Figshare, I got to know the ROpenSci team and I could bravely say: “Open data has no cost: it is time investment done at the right time”. Indeed, it occurred to me that I had saved time by getting a workflow aiming at open data per default. For instance, while the manuscript was ready, I decided to add a violinplot over the boxplot for all figures (9 plots for 3 figures): it took me about 15 minutes from the time I had the idea to the moment the figures were inserted in the overleaf produced manuscript. The same process with a “normal” workflow would have taken at least two days.

Open data has no cost: it is time investment done at the right time

You’ve also recently built a ScienceOpen Collection on Alzheimer’s Disease. Why did you choose to do this, and how does it fit into your current research plans?

I got recently involved in a project aiming at curating the preclinical literature for behavioural tests used to model specific neurological diseases. As a pilot, we are analysing one mouse model of Alzheimer’s disease. I thought to use the collection system of ScienceOpen to:

1. Gather the literature on one place;
2. Having a platform to discuss putative discrepancies between results and hypotheses;
3. Be sure that this work will not end in a drawer in cases there is nothing worth a publication in our findings.

Where do you see the future of scholarly communication? What steps are needed to get there? And whose responsibility do you think it is to lead this change?

I do not care about scholarly communication much. My thing is data quality and science reproducibility. By doing the plan about data archive, analysis and publication before starting to gather data, we could get rid of most problems we have in science reproducibility. As a plus, we would also be faster, and as a by-product, scholarly communication will change, we will see more micro-publications and data-publications for instance.

If you could give one piece of advice to students looking to pursue a research career, what would it be?

First, be sure about what you want: if your dream is doing science, do it without thinking about making a career. If your dream is having a research career, do it without thinking about science. For a successful career, you have to make your move in the 3 first year after your PhD, and get a 5 years grant during your 4^th year. If you miss that train, give up the career path and try to do science, inside or outside of the academic universe. But above all: have fun and get to know as many people as you can.

If your dream is doing science, do it without thinking about making a career. If your dream is having a research career, do it without thinking about science

Thanks for your insight Julien, and all the best in the future for Drososhare and your career!

Julien is currently a postdoc in Neurobiology in Berlin at the Charite. He is the founder and CEO of Drososahre, a company for automating fly sharing. In Berlin, he also organises the Open Science Meetup. His primary research interests include understanding learning behaviours using fruit flies as a model organism. He started his higher education in Fribourg, Switzerland, where he received his PhD in 2006.