Science 2.0/Brainstorming

This page will be a discussion on applying the principles of Web 2.0 to scientific communication. I wrote up this page just to get the conversation going, please add categories (or rename it :). Kathleen's open science comments from her soapbox are a good starting place as well. Also, our previous discussion on future approaches to publishing.

Science 1.0->2.0

Web 1.0 and Web 2.0

The internet is undergoing a major change - from an original environment where individuals posted static information that was hard to navigate to a new environment where people are dynamically posting information and collaborating. New search and aggregation tools are making it easier to find and contribute to the information that an individual is interested in. This shift has been described as the switch from Web 1.0 to Web 2.0. Some examples of old and new web technologies taken from an O'Reilly Web 2.0 articleare listed below.

• Web1.0-->Web2.0
• Britannica Online-->Wikipedia
• directories (taxonomy)-->tagging ("folksonomy")
• mp3.com-->Napster
• Akamai-->BitTorrent

Examples in science

Science can both use these new tools directly and also apply similar concepts by analogy. The expectation is that these new concepts will make the biology community more open and will increase the speed that new information is accumulated and disseminated. Listed below are some of the ways that science is already moving from Science 1.0 to 2.0. Later we will discuss some new ways that scientific research could be accelerated using new communication and collaboration tools.

• Regularly scheduled printing of journal issues --> Continuous release of articles in online format.
  • Example: Nature --> Molecular Systems Biology
• Subscription-based journals --> Open access journals
  • Example: Nature --> Molecular Systems Biology
• Peer-reviewed specialty journals --> Articles aggregated and ranked by search engines (Google style) or via a catalog and user review (Amazon style)
  • Example: Pubmed --> Faculty of 1000.
    Faculty of 1000 gathers papers by topic and allows people to post evaluation to them. It does not appear to be open access.
• Maintenance of static protocols in labs which are shared via reference texts -> Online sharing of protocols
  • Example: Individuals' and labs' private protocols --> Protocols Online
    People post protocols, others can post reviews of those protocols.
  • Example: Molecular Cloning --> OpenWetWare protocols
    OpenWetWare permits anyone to share, edit and annotate protocols.
• Methods & Techniques Publishing --> Sharing materials
  • Example: Methods papers --> Caltech's foundry for microfluidic chips
    • Out of curiosity, when you design a chip to be fabricated by the Caltech foundry, do they retain your design? Do they share it with other users? My question is really to what extent do users of the resource contribute to the resource in this example.~ cmc 12:33, 10 Nov 2005 (EST).
• Materials collections --> Materials registries
  • Example: Strain collections --> MIT Registry of Standard Biological Parts
    The Registry permits users to design, obtain, test, and annotate parts. The users who design parts do not necessarily have to be the ones who synthesize the parts. It enables anyone to post comments about any part. Eventually the Registry would like to move to a web of registries based in multiple locations to further decentralize and make the engineering of new and existing parts easier.

Drawbacks of the current situation

There are drawbacks to the current approach to scientific research.

1. Current publishing approach is very slow: results in yesterday's journal are 0.5-2 years old.
2. There is only one main level of information dissemination: peer-reviewed, highly polished summaries of work in journals.
   • There doesn’t seem to be a good reason to have only one form information distribution in science. While it is important to have tight, very trustworthy summaries of research it could also be useful to have “streaming results” from a lab updated weekly or monthly. Understandably, readers would need to take these results with more skepticism than those published that month in Nature; however, it doesn’t mean they would have no value.
   • In the past, conferences served as the primary means for rapid interaction between groups. However, the current atmosphere makes it such that people often only talk about published results. Much of this phenomenon has to do with the structures around competition for funding sources.
3. Current publishing approaches do not encourage open feedback and reviews of work.
4. Published articles often do not contain sufficient information to reproduce the results of the work. For instance, compare a two paragraph Methods section to having access to a complete lab notebook.
5. Journals articles are edited and formatted for a print medium while being distributed primarily online. In other words, they fail to take advantage of the features offered by the web.
   • The near absence of hyperlinks in journal articles is the clearest example of this problem.
6. Published data is often not made available in a machine-understandable format.
   • Oftentimes the data from scientific research is not necessarily published in a form that encourages other groups to take the data and analyze it independently. Presenting work in a form that enables others to make use of it easily would enhance the quality of the work. Such goals are driving the scientific semantic web and SBML communities.

Ideas for future Science2.0 projects

• Slashdot for scientific articles, ideas
• Online lab notebooks
  • Accessible from everywhere
  • Searchable
  • Referencable
  • Equipment-friendly: as data comes off a machine, it goes directly into notebook.
• Network- or internet-friendly lab equipment
  • All lab equipment actually becomes operable over the web and therefore does not require a dedicated computer to run. (Note that Randy is actually considering buying a robot that is programmed on a computer and transferred to it via a memory card.)
  • Permits sharing of equipment across the world opening up science to a wider rangle of people.
  • Requires standardization of communication protocols.
• Virtual posters online
  • Currently, researchers still rely on static posters to describe their work at conferences. Instead, posters culd be used as an overview and advertisement for a fuller description/discussion of a project on OpenWetWare rather than an end in themselves. At conferences, attendees could read a poster, visit the OpenWetWare page, and then return to the poster for a fuller discussion with the author in person or online.
• Lab "feed"
  • An equivalent of an RSS feed of what is coming out of a lab updated daily/weekly. Results would be less finalized but might help people coordinate on projects accross labs rather than just repeating each other's work in secret.
• Immediate data sharing
  • Currently authors must keep data private until a paper is published. Regardless of whether the publishing process can be accelerated, it would be desirable if the raw data could be shared earlier while protecting the authors rights to publish papers based on the data. This idea is related to some of the suggestion/problems above.
• Scientific "currency" outside of authorship on paper
  • I think this might enable better data sharing among other things. I.e. if you use my data you need to give me X credits (where the value of X credits to my career is less than authorship on paper but greater than nothing.) I have no idea what such a system would look like.
• Collaborative written works
  • Currently review articles written by experts in the field are the primary mechanism in which the state of a research area is evaluated. Instead, one could imagine using collaborative writing tools like wiki's to maintain a real-time synopsis of a field.

OWW as an ongoing 'Science 2.0' experiment

OpenWetWare is good example of Web2.0 meets science. Even though it hasn't been around long it already has provided some "in the trenches"-type information about where Web2.0 will mesh with traditional scientific research approaches and where it will raise concerns.

Content generation

Lowering the barriers to content generation and dissemination leads to more informal discussion and more up to date results. Most of the information on OpenWetWare are things that would have been archived anyways, just not in a centralized stored format which is both open access and open editable.

Meshing

• Standardization and sharing of protocols. See DNA Ligation.
• Makes information more accessible. Not only is my information (i.e. lab notebook) accessible to others but it is accessible to me from anywhere. See online lab notebook above.

Common concerns

Trustability

People often are worried about how trustable the people on OpenWetWare are. We often sidestep this by saying everyone has a login, and we can track any change anybody makes. When the first 'vandalism' occurs, we can rethink what our policies of once you are logged in, you can edit everything. However, at some point, vandalism will occur. What do we do then? How do we convince people there is way more good, than the occasional bad.

Getting 'scooped'

When we try and promote OpenWetWare to people one of the most common responses is “if I post my project details there I will be scooped.” We usually ask that they consider a few scenarios:

1. The project is shared online and someone steals the idea and publishes before you without ever mentioning they are working on the project. Publications (science currency) = 0.
2. The project is shared online, several people email you saying it looks cool, you push those connections and establish 4 new research directions. One of which is even more productive than the original direction. Also, a competitor steals one of the ideas and beats you to the punch. Science currency = 4-1 = 3.
3. You keep information to yourself and publish your work. Science currency = 1.

What this example illustrates is that the question is whether you think the probability of scenario P(2)/P(1) > 1. This also approximates one of the Web 2.0 principles “the service gets better the more people use it,” since the more people that buy into this model the higher P(2)/P(1) gets. We think P(2)/P(1) is already > 1, however we’re naïve graduate students. However, if we're right than ‘natural selection’ will lead those adopting an open approach to science to generate more science currency and thereby outcompete those taking approach (3).

Looking unprofessional

There is a concern that "posting non-polished, uncompleted research/documents online will make the lab appear unprofessional." Granted that some wiki pages are quite obviously under development and really serve as a tool for brainstorming. However, other wiki pages have gone through more extensive review and present a reasonably cohesive explanation of a topic. The disadvantage of having some under development pages is likely outweighed by having a rapidly updated space that makes the lab appear alive and dynamic. Anecdotally, some OpenWetWare users have been approached at conferences by people who have seen work on the wiki and the response was pretty positive. Similarly, post-doc applications to the Endy lab have increased since OpenWetWare began.

Others

Other common concerns (permissions, vandalism, "own space") provided here