Open Source, Open Data, Open Access, Open Notebook - Open Science Summit 2011 challenges patents, research methodology
The Computer History Museum in Mountain View, CA was abuzz with coffee-fueled conversation the weekend of Oct. 21-22nd as a very diverse bunch - scientists, biotechnology lawyers, software designers, students, hardware developers, bioinformaticists, futurists - gathered at the second annual Open Science Summit for a two-day powwow.
The venue was aptly picked for the conference: installations honoring the contributions of many Open-Source greats such as Linux developer Linus Torvalds suggested many parallels that can be drawn between the conditions that spurred the explosive evolution of the computing technology twenty years ago to what may now happen in other tech fields such as biotech.
Future of Innovation
While the concept of Open Source is intuitive for software...what exactly does it mean when applied to science? The goal of the movement, and of the Summit specifically is to address methods which would transform the way we do science, making it radically more effective. “We have the chance to adopt new models and build new institutions better suited to conducting distributed, massively data driven, collaborative science in the 21st century,” explains conference organizer and open science proponent Joseph Jackson. Reforms across many financial, legislative, and technological facets of the current academic and corporate research industry have already begun to redefine innovation. Talks ranging on topics such as the future of Peer review, Patents and IP, scientific funding mechanism reforms, and the emergence of the biotech hackerspace movement addressed these budding trends. Below are some notable outtakes that addressed the fields of innovation and IP:
Open Source, Open Data
The conference kicked off with an eye-opening (albeit the early hour) discussion of Transparency in research by Victoria Stodden, a Columbia University statistics professor who serves on NSF's Advisory Committee for Cyberinfrastructure. There is an absence of data and code sharing in the current peer-review publication culture, with only 20% of computationally-oriented articles providing the code package. This is a problem, the speaker argued, of reproducibility and results verification, that exists at the level of the scientific method itself. And it's becoming more exacerbated as science is becoming more computational in nature. “You know how hard it is to replicate code based on a 4 page description of their result,” said Stodden, alluding to the fact that computational results may not be getting cross-verified by the scientific community.
Audience reactions to the talk reverberated in real-time through twitter (#opensciencesummit). “A potential 3rd branch of scientific method, after hypothetico-deductive and empirical: computational” tweets one attendee. It seems that while we had hundreds of years to think about the first two branches, designing the computational branch is something the community will have to do now.
Moreover, there is another hurdle: conflicts of interest and intellectual property. Scientists wouldn't necessarily want to make their results easily reproducible out of fear of getting scooped. They may also feel like they are onto something they can patent and start a company. This poses a cultural issue that undermines reproducibility, a “long-term norm” in the scientific community, argues Stodden. Is there a way to encourage scientists to not fear opening up their code and data? Perhaps there is a way, proposes one attendee during the Q&A section, suggesting that a methodology is devised for attributing authorship for code each time it's reused. Another way to safely share data is though initatives such as the web application Collaborative Drug Discovery, presented by speaker Barry Bunin. CDD is an online datasharing program, that allows scientists to browse for specific research topics and find datasets published by peers.
The Patent Game
The current patent system got quite an ear-pulling in the talk block titled The Future (the End?) of Intellectual Property. An especially interesting talk featured the 'Patent Game', a video game that simulated how well the current patent system actually stimulates innovation. In this game players competed to create inventions and patent them in order to maximize profits and win a cash prize. “Before i sold my soul to become a lawyer I was a geneticist,” jokingly begins Andrew Torrance, a biotechnology patent attourney who, in collaboration with the Social Code Group at UC Irvine, headed the study. While the patent as a tool for promoting “progress of useful arts” is essentially written into the US Constitution, practically no empirical evidence exists that without patents you will get fewer innovations. Torrance was on the hunt for this evidence.
The “Patent Game” allowed players to combine “elements” to make inventions. A player could make things, then sell on market, or patent first and then later make and sell. Prior art nullified the rights to get patent. Additionally, players could buy, sell, licence rights, and hire lawyers to sue each other for infringement. Real court data was used to simulate court damages. Additionally, litigation affected one's inventing abilities in terms of both money and time. When comparing worlds with “patent innovations”, “commons”, or both, data demonstrated that the “commons”, aka open-innovation world did significantly better in terms of unique inventions per second scoring 100 unique inventions per second compared to the patent world's 80 per second. The “commons” paradigm even more dramatically outcompeted the patent innovation paradigm in terms of productivity (measurement of replica and unique inventions) demonstrating a whopping 60 invention per second lead on both of the other systems as well as the wealth per capita measurement, where the purely open-source “commons” system similarly outperformed the other two.
Is this result really all that shocking?
While conventional belief is that open innovation will result in sub-par products because competition will destroy each other, this point of view is currently facing some hefty challenges. This idea was first challenged by economist Elinor Ostrom, who received the 2009 Nobel Prize in Economics for her work in economic governance, which demonstrated that purely social (in contrast to legal) mechanisms can be used to successfully manage common resources without government regulation, or privatization.
Biotechnology law especially is fond of the current patent system, however this is challenged by some heavy opposition [1]. The Patent Game results are certainly not the only evidence to challenge conventional wisdom. There is mounting evidence that the current intellectual property law needs major reforms in order to optimize science innovation specifically [2]. Nevertheless, as Mr. Torrance put it, today Open Science remains the “Eccentric Cousin” of the biotech field, though that certainly is beginning to change.
Crowdsourcing
Crowdsourcing and crowdfunding also received significant attention at the conference. Crowdsourcing, or the application of online forum-based collaborative problem solving to science, first gained mass attention when the Cambridge University mathematician and Fields medalist Timothy Gowers challenged the readers of his blog to write a general proof of the so-called density Hales-Jewett theorem. The problem was solved in six weeks, and published under a collective pseudonym. Optimizing the interactive math problem-solving approach has been the key interest of UC Berkeley graduate student and Open Science Summit speaker Anton Geraschenko, who founded the math question repository MathOverflow.net.
The concept has other success stories, notably crowdsourcing the study of an eColi outbreak in Germany, discussed at the Open Science Summit by Beijing Genomics Institute scientist Joyce Peng. It has also created a flurry of biotechnology games. For instance, the players of online protein-folding game Foldit were able to in as little as three weeks make a 3-D model of an especially tricky HIV enzyme[3].
Talks by Dr. Ingmar Riedel-Kruse (Stanford University) and Dr. Rhiju Das (Carnegie-Mellon University), co-authors of the game EteRNA discussed the broader impact of biotechnology games. The ability to harness the pattern recognition and ingenuity of many human participants to solve real-world problems has become a compelling approach since it gives better results than the most elaborate algorithms to date. EteRNA is the successor to FoldIt in which non-scientist players design new ways to fold RNA molecules. Each week, the best designs are synthesized at Stanford University, and tested for proper folding and biological activity. In the future, they may be combined into a nanoengineering toolkit of RNA that, once transcribed in a cell, may do things such as sense light.
Video games are also a useful way to teach, not just solve. In his talk, the CEO of Primer Games Alex Peake talks about his experience with Code Hero -a first-person shooter game that teaches you how to program. In this game, you find yourself in the 3D world of Codia, where, upon arriving to a place called API, (the “Hogwarts” of Code Hero) you meet none other than Ada Lovelace, the world's first computer programmer who becomes your coding mentor. Throughout the game you learn how to use a 'code gun' to literally shoot java code and alter the gaming environment around you.
Finally, crowdsourcing is also used to direct growth of innovation by crowdfunding compelling projects. In his talk, Jai Rangathan presents the SciFund Challenge, an initiative that takes the idea of crowdfunding through online donation mediators such as kickstarter.com, and tailors it specifically to science by attracting key research publications.
Open Notebook, Open Genetics, Clinical Trials 2.0
Re-organizing medical research, optimizing patient-expert communication, data-driven drug repurposing, optimized electronic lab organizers, social networking sites that link reserachers and allow scientists to outsource experiments - these were all topics of discussion at the Summit, with researcher panels providing forums for discussion.
A budding field of open-source medical and scientific data management and datasharing start-ups presented their goods. With the cost and speed of sequencing decreasing dramatically, the ability to have one's DNA sequenced quickly and cheaply is virtually here. Therefore, the next generation of personal genomics, genomic regulation, mass genomic data mining and interpretation attracted a large panel of speakers and start-ups.
Open Science Summit and the hacker tradition
All in all, the Open Science Summit is not your cookie-cutter conference: smartphone-driven camera carrying robots invade the conference room, while conversation buzzes around the book signing of 100+, an analysis of longevity by futurist Sonia Arrison. The local hackerspace for biotech, Biocurious, collects emails and gives out flyers, while eager young engineers demo their new haptic-feedback gaming devices or prototype augmented reality headsets. The gender and age gap is quite noticeably blurred, and the playful, innovative hacker spirit is clearly present among the speakers as well as attendees. Engineers, frosty-haired emeritus scientists, lawyers, enterpreneurs, doctors, futurists, longevity enthusiasts, blue-haired videogame designers, and hackerspace activisits demonstrate the malleability and fusion of ideas that is required for successful innovation. These ideas promise to re-write and optimize the rules by which we do science. Make no mistake, the Open Science movement is the stuff of technological change.
Conference footage is available here.