Difference between revisions of "Open collaborative design/Open Source Scientific Research"

Revision as of 13:24, 1 September 2010

Scientific research can also be done in a spirit of open collaboration. Scientific papers can be made available online to be read and reviewed (a process known as open publiushing), and the raw data collected from scientific studies can also be made publicly available (known as open notebook science).

The main platform for the scientific community in the 20th century was the peer-reviewed scientific journal. A scientist writes a paper, submits it for review by other scientists and, if it is found to be up-to-standard, the paper is published in a copyrighted journal and made available to subscribers.

This model has many advantages: the main one being that poorly conducted experiments are filtered out by the peer-review process. On the other hand, subscription to journals is often expensive (e.g. It costs $200 a year to subscribe to Nature), limiting access to scientific knowledge. Negative results (e.g. finding that a certain drug has no effect on a disease) are rarely published in such journals, as space in the journal is limited and tends to be given to more interesting positive results.

An alternative to the peer-reviewed journal is open scientific research. This means posting scientific papers freely on the Internet for anyone to access, without intellectual property restrictions. A different kind of peer-review then takes place, an open peer-review like what we see on Wikipedia. Other scientists can rate the paper or comment on it, allowing the best research to rise to the top. Data-mining algorithms can link related research studies together, showing studies that use similar methodologies together, even aggregating the results of several studies.

Open collaboration in science allows scientific experiments to be global collaborations of interested parties around the world. This allows for greater n numbers, as data can be aggregated from a large number of researchers. This leads to more reliable results.

When data and methodologies ae published openly, anyone can point out mistakes or oversights. This allows them to be swiftly corrected.

A key tenet of the scientific method is that all experiments be repeatable. When anyone can read about an experiment and replicate it for themselves, false results - due to sloppy design, fluke or fraud - are neutralized. By making scientific papers open, experiments can be repeated by anyone, not just subscribers. This will ultimately lead to more reliable scientific results.

Another application of open collaboration to scientific research is allowing people to contribute processing power to computing for scientific research. Rather than have a supercomputer crunching numbers in a lab, thousands of people's personal computers around the world can be used. This has been very successful at helping calculate how proteins fold — see http://folding.stanford.edu. The Open Science Grid is a cloud supercomputer used for data-rich research.

There are already millions of papers and datasets contributed to open scientific platforms. GenBank is a public library of over 120 million DNA nucleotide sequences and the proteins they produce. The Public Library of Science publishes open scientific journals in eight different areas. The Human Brain Project is a huge collaborative project that aims to share all known data about the human brain.

Links

• http://sciencecommons.org
• DIY Bio 
• http://openwetware.org
• H+ Open-source science article - Joseph Jackson interview
• Scientists Embrace Openness - an article by Chelsea Wald
• http://www.biomedcentral.com/