Git for science: SciGit!

Screen Shot 2013-09-20 at 4.53.16 PMMost scientists will tell you that collaboration is key research. Virtually all research projects consist of associations of different persons with complementary skills (+ that one parasitic PI with a high-up position ūüôā ). Collaborating with your labmates is one thing, but larger scale or long distance collaborations require significant attention to track, organize, and synchronize the work. Collaboration tools for researchers can help greatly and online tools to organize experiments or share data efficiently are becoming more popular.

But when comes the times to write a publication, collaborations typically get very messy. Files names become overly complicated as the version number increases; who hasn’t had a TC_v17.1_Final2 at the end of their word documents? Files are exchanged by email, do not get send because of the size, get forgotten in the inbox or spam, and versions get swapped. Luckily SciGit, a version control tool for publication writing has recently went live! Inspired by the hugely success GitHub, the tool advertises three essential functionalities:

  1. Collaborate without mess.
  2. Supports your editor.
  3. Safe and secure.

In sum, with SciGit you can easily track document versions and easily visualize edits while continuing to use your usual off-line text editor such as Microsoft Word. The modifications to the documents can be tracked either online or through the free desktop client (only available for windows at the moment). Deletions are simply highlighted in red and additions in green. The documents are automatically uploaded to the cloud in a safe and secured location with tight control over who can view the documents.

The tools is still in beta and seems to have¬†space to evolve. But keep and open mind and give it a try! Comments are welcome ūüôā

Runmycode gives a second life to your code

Logo_RunMyCodeThe web 2.0 technologies can help generate new science, but can also greatly help after publication. For example, some of you are spending months or even years coding the perfect model for your study. You will input your parameters find interesting trends, publish and move on to another project.

But then what? Don’t you feel your work is left partially untapped?

What if your work could be applied to slightly different conditions or adapted to other problems? What if users from around the globe could access that code, and run it directly online?

Runmycode does exactly that. It allows you to create a companion website for your publication. The site can host a cloud-based version of your code that users can run at will. The users input the parameter, press run and generate new data. It’s that simple!

Screen Shot 2013-02-08 at 4.08.53 PM

Runmycode is a non-profit¬†initiative. It resulted from a collaboration between many¬†individuals¬†and¬†institutions,¬†mostly¬†located¬†in France. I believe that this sort of “extension”¬†to the classic publication is exactly what science needs to be more open and better communicated. Now only if a “RunMyExperiment” website for biological research could come-up…

Online tools that help you share your data.

One important aspect of the open science movement, is making your research data more accessible than when formated in manuscripts (which are usually only the tip of an enormous iceberg of work and findings). The data can be slide presentations, data sets or figures for example. All are very rich in unpublished information. Several sites offer researchers to share such information:

I can see a number of different advantages to doing this.

  • Online sharing services help create a communities and form relationships.
  • Sharing protocols or workflows increases research efficiency, by favoring access to ¬†expertise and avoiding reinvention.
  • Platforms for sharing data allow to share negative results and help combat the “positive result bias” in scientific publishing.
  • Finally, this is also increases your presence on the web for your¬†personal¬†promotion.

Of course in some scientific domains, sharing openly your data is discouraged because of strong competition or pressure to patent or publish first. But even if posted online years after the original publication, the data can still be valuable to the community.

Searching the pool of shared data is currently not an easy task. Although sites like SlideShare, OpenWetWare and FigShare are rather well indexed in by search engines, searching for specific information contained in the data might be more challenging for example. More specialized tools, like what the DataCite initative is offering,  that can efficiently index, organize and search through the shared data will clearly benefit the open data movement.