Sunday, September 28, 2008

Open Notebook Science Challenge

The wiki for the Open Notebook Science Challenge that I proposed during my UK trip is now available. We are currently looking for sponsors and participants.

Open Notebook Science Challenge


The first round of this challenge calls upon groups or individuals with access to materials and equipment to measure the solubility of compounds in organic solvents and report their findings using Open Notebook Science .


Understanding exactly how an experiment was performed is essential to the efficient progress of science. There are no absolute facts in the scientific literature; every measurement reported is only meaningful within the full context of how it was generated. The purpose of a laboratory notebook is to report as much of this context as is reasonable. But to find trends data must be abstracted to a level where they can be manipulated in tables and charts. This is not a problem as long as one can drill down from each data point in a chart to the full context found in the laboratory notebook.

For example, a Google search for "vanillin solubility in THF" pulls up a lab book page EXP207 where it is reported to be 3.89M. This number might be used in a table of someone trying to quantify trends or test a mathematical model, in which case reliability of the number is important. By reading the lab notebook page it becomes clear that 118.5 mg of solid was measured on a scale with 0.1 mg accuracy. However only one measurement was obtained. All kinds of other details which might be important are provided, for example how long the mixture was vortexed, at what temperature and the physical appearance after evaporation. If this number turns out to be an outlier, one can investigate if a calculation error was the cause by inspecting the linked spreadsheet.

However, if a researcher is simply looking for the feasibility of making up a 2M solution of vanillin in THF for a reaction the margin of acceptable error is so wide that the answer is almost certainly "yes".

The purpose of Open Notebook Science is to allow immediate communication of scientific results. The value of these results will depend upon the quality of the laboratory notebook and the linked raw data. Publication in peer-reviewed journals is still an extremely important part of this process but it is not an appropriate vehicle for the efficient communication of this type of information.

In fact, one of the motivations for participating in this project is that we will collect data from sufficiently well recorded experiments and publish them in a peer-reviewed journal with the participation of the researchers as co-authors. We aim to build a mathematical model to predict solubility using the results obtained from this project.



Jean-Claude Bradley
Cameron Neylon
Rajarshi Guha (modeling)


Simply request an account on this wiki and start recording experiments using a format similar to UC-EXP207 . The organizers will provide feedback in the form of comments in bold and italics directly on the wiki. Hitting the Recent Changes link on the left navigation bar is a good way to keep track of edits.



Labels: , , , ,


At 10:54 AM, Anonymous Anonymous said...

Is there a centralized service for providing open notebooks to scientists without the tech savy/time to set one up themselves?

At 12:22 PM, Blogger Jean-Claude Bradley said...

For this particular project participants just need to request an account on the wiki, which serves as the actual lab notebook. Images can be stored on Flickr or on the wiki. Short videos can be stored on YouTube. Tabular information can be stored on GoogleDocs. All of these services are free, hosted and pretty intuitive but if anyone runs into a problem I would be happy to assist.

At 3:22 AM, Blogger Cameron Neylon said...

Kris, we do keep talking about doing this but no-one yet has had the time to set something up comprehensive. In the meantime as Jean-Claude says we are mor than happy to help people set up and advise on what system(s) are likely to be best for them.

At 10:26 AM, Blogger Kris said...

Jean-Claude and Cameron,

I am the kind of person that would love to create a comprehensive online lab notebook especially since it does not exist. I would like to get input from you 2 and anyone else that is interested as what features you think would be the most valuable in such a system.

At 12:07 PM, Blogger Jean-Claude Bradley said...

Speaking just for my lab, at this point in time, a wiki really serves our needs quite well. When we get into more automation that might change. For this challenge using Wikispaces for the actual lab notebook will work fine.

At 9:27 AM, Anonymous Anonymous said...

Kris, I would agree with Jean-Claude that Wikispaces is a good place to start. If you want a few more features then take a look at as well. Its a good idea to play around to start with and see what works for you. What kind of area do you work in?


Post a Comment

<< Home

Creative Commons Attribution Share-Alike 2.5 License