The case for open science
June 17th, 2007Traditional science is carried out by experimentation, interpreting the results in respect to the hypothesis, and repeating this until publication. Publication is everything, and as such it’s important to keep your research secret, lest anyone publish something similar which undermines the opportunity to publish in a high impact journal.
An open science philosophy says that this closed approach slows scientific advancement. Results should be made available as soon as possible, then everyone in the field can benefit from them sooner. There is no delay for peer review and publication.
Why you should participate in open science
A pure motivation is sharing knowledge for knowledge’s sake. Other scientists get access to your research faster. This can prevent groups working on overlapping topics, or research that would be invalidated by new results. Both are possible scenarios, given the time between discovery and publication.
There are also more selfish benefits of open science. Scientists are always more than willing to give their opinion. Imagine you discussed on your website a hypothesis about a gene, and you planned to test it by expressing it in E.coli. Then another researcher, reading your site, wrote to tell you that they had tried to the same but had been unable to. This would save you a lot of time by preventing you from repeating the same mistake. This isn’t a hypothetical situation though, this is exactly what happened on Rosie Redfield’s research blog. This is just one scenario. People willing to tell you a better way to perform an experiment, or alternative interpretations of results are other possible benefits for the researcher willing to publicize their results and ideas early.
Why you should shun open science
There’s one very obvious reason for being cautious about making your results public before you publish - will people steal your research for their own benefit? Unfortunately the answer to that is likely to be yes. While we would like to think that everyone would be honest enough to acknowledge your work if they use it, some won’t. The nature of the scientific game means that we get points for publishing in important journals, and the more points we get the better the scientist we’re regarded as. If someone can increase the significance of their work by repeating your analysis, as their own, then there’s always the possibility they might. I don’t think we should be surprised at this, since this is the way the system is set up, and until we can come up with a better one this always likely to happen.
There are also other people that have a stake in whether you make your research publicly available. Your work is likely sponsored by a research council or funding body. Your institution gives you a place to carry out your work. Co-researchers contribute time and data to the project. All have something to lose if you’re risky with your research, and are scooped. They all may be less willing to work with you as a result.
There is one last point to consider. If you make public a figure that describes your research, that figure may then be ineligible for publication - similarly for conclusions, and interpretation of data. Journals have different policies on what researchers can make available prior to publication. It’s worth taking a moment to consider this before putting data on a website.
A middle way
Despite all this I am very in favour of open science, and would like to make all my research available for any one who is interested. However the reasons against are considerable, and should not be ignored. But after speaking to other people in my department, we’ve come up with a middle way.
There are always experiments that don’t work out. The results were inconclusive, I have no idea of how to interpret the results, or the experiment just didn’t work. I’ve got nothing to lose by making this data publicly available, I’m not going anywhere with them anyway. However, that’s not to say they might not be useful to someone else, or even better somebody could give me their opinion on where I’ve gone wrong. I think this could be a good start for my experimentation with open science.
June 19th, 2007 at 9:23 am
Best of luck :). Another possibility is also to share directions that you do not intend on pursuing. This is what I have been trying so far. Sharing ongoing research for me is hard because must of the stuff that I have been working recently have been collaborations and it is hard to talk several people into sharing results online (specially when it is a naive PhD student trying to convince them).
June 20th, 2007 at 4:16 pm
Thanks Pedro
Could it be that open science is inevitable, it’s just how long it takes to get there? The concept is definitely gaining more and more ground. It could move slowly through the research community, or a really good example it could convince a lot of people. Probably both.
I’m sure blogging positively encourages open science too. Maybe it’s part of the state of mind.
June 21st, 2007 at 6:56 am
I would really love to see a more open approach like the one you describe. As I see in the field of microarrays, the jealousy and the secrets surrounding the research sometimes are even against common sense. And sometimes secrecy is not worth it as the published results leave a lot to be desired.
I believe that the concept of “publish or perish” needs an overhaul… right now it almost feels each lab is a secret society working on some (secret) plan for world domination.
June 21st, 2007 at 9:16 pm
[...] the theme of open science, I think one of the most important issues that needs clarification are the journals position to you [...]
June 21st, 2007 at 9:41 pm
Thanks Luca. I think your example of microarrays is a particularly good one. There are a lot of microarray analysis tools being published. I think that this leads to researchers suffering when they need to analyse their data, as there are a huge selection to choose from - it’s difficult to know which one to pick. As you say, the publish or perish philosophy encourages bioinformaticians to create and publish a new microarray tool, rather than improve an already existing one. The latter option is the one more beneficial to the community. A related result, a quick survey carried out by Stew showed that 12% of applications published in bioinformatics one year ago were no longer available.
June 22nd, 2007 at 9:19 pm
Mike, I can relate to that finding by experience. Over the three years I’ve been working in my group I’ve tested several microarray tools. Most of those (but luckily, not all of them) had been abandoned along the way, or had bugs that prevented proper use. Some even (which IMO is somewhat shameful) presented algorithms without proper implementations, which were of course impossible to test.
Recently I was at a meeting over an improvement of a pathway analysis tool and the software engineer who wrote it said something like “We always reinvent the wheel because other groups don’t share information”.
For these cases I advocate the use of licenses such as the GPL for biological software (or even BSD) so that at least someone else can pick an abandoned project and improve it.
June 24th, 2007 at 3:09 pm
[...] Barton at Bioinformatics Zen tells us his opinion on Science 2.0. I loved this sentence: Despite all this I am very in favour of open science, and would like to [...]
June 24th, 2007 at 8:48 pm
Great topic!
Thank you Mike!
I think it’s will be very useful for everybody who read this topic if we can give some best examples of Open Science projects,
in my minds -
http://www.plos.org
http://www.biomedcentral.com
http://precedings.nature.com
June 25th, 2007 at 1:41 pm
@ Luca
It is a shame that bioinformatics tools recieve little support once they are published. I think the scientific system does encourage this - once the tool has been published there is little reward for sustaining support. Also people leave labs, and funding runs out for projects which means that there is no able to support the tool.
As an extreme example imagine a case where instead of contributing code to the linux codebase, everybody created their own distribution to include the feature they wanted. I think this something like the scenario you are describing. Perhaps if the journals encouraged tools to submitted with a GPL license, as well as papers that were based on improving already existing tools with new functionality.
@Nanog
Thanks for your comments.
I’m working another special edition for bioblogs, the focus on open science. You’re more than welcome to take and look and contribute.
June 26th, 2007 at 10:19 am
Mike, you might be interested in similar musings from Frank Gibson at Newcastle:
http://peanutbutter.wordpress.com/2007/06/26/do-scientists-really-believe-in-open-science/
September 10th, 2007 at 6:06 pm
[...] still very keen on the open notebook science movement: I think open collaboration will benefit everyone involved. So, when I should have been working on the Bio::Blogs special edition, I was instead tinkering [...]