When Peking University biophysicist He Jiankui recently announced that he had genetically altered human embryos that were subsequently taken to term and born as twin girls, he was immediately denounced by huge portions of the international scientific community. While much of this criticism stemmed from the ethical quandaries posed by He’s actions, an even more fundamental reason for outrage emerged from the fact that he had worked in isolation. Like some rogue operator in a science fiction movie, He apparently pursued an elaborate research program that seems to have been utterly unknown to others in his field, a violation that runs even deeper than any breach of established guidelines for the use of genome-editing techniques.
This commitment to openness and transparency is taken to be a hallmark of the modern scientific enterprise, as the widespread condemnation of He reveals. Scientists meticulously itemize and explain their work to highly critical audiences, who will quickly and thoroughly seize upon any aspect that appears questionable. Details abound in every scientific publication, outlining each step in the research process.
Yet even in the midst of this wealth of information, some researchers are calling for more. They point to the fact that access to raw research data is regularly restricted, a move that might be regarded as a means of safeguarding the integrity of what may be very complex databases. Such restrictions also anticipate proprietary application of data, so that if a set of scientific findings becomes the patentable basis of a commercial venture, the flow of information will be further limited.
That prospect does not sit well with many observers, some of whom advocate loudly for an entirely open approach to science. A great deal of research is conducted with public money, they argue, which makes the recipients entirely accountable for revealing everything that was done with taxpayer money. There is also a case for making privately supported science as accessible as possible, based on the possibility that a locked archive could effectively hide productive insights or practical solutions from external investigators.
“Open science is a tall order — how one collaborates across the globe and shares data,” says Mario Pinto, former president of the Natural Sciences and Engineering Research Council, who discussed the social responsibilities of knowledge sharing at Simon Fraser University this fall. “Certain communities have done this extremely well, such as sub-atomic particle physicists or genome scientists. That was initially because of the high costs of doing research, but the culture has pervaded what they do even as costs have come down.”
He acknowledges that chemistry is among the disciplines where there is still some progress to be made, but he also points to models of open collaboration that are showing the way forward. Among the most prominent of these is the Structural Genomics Consortium (SGC), a not-for-profit organization founded in 2004 and with research hubs at 7 universities around the world, including McGill and the University of Toronto. Participants make all their research results freely available, such as an online database containing more than 1,500 high-resolution protein structures, and chemical probes, which are crucial to pinning down the relevance of prospective drug targets.
According to SGC CEO and Director Aled Edwards, a professor at the U of T, the development of chemical probes was traditionally considered the remit of pharmaceutical companies, were almost always patented, and were shared only under highly restrictive terms. Our decision to make research tools freely available removes a central obstacle that has been blocking drug discovery for decades, and enables academia to accelerate discoveries. Already more than 50 clinical trials are in progress based on discoveries made around the world using SGC’s open chemical probes.
The potential of open science for accelerating translational impact is driving a parallel initiative at the Montreal Neurological Institute (MNI), which was founded in the 1930s by one of the world’s leading brain researchers, Wilder Penfield. His pioneering efforts set the stage for huge leaps in our understanding of the nervous system and diseases affecting it, but current MNI director Guy Rouleau points out that there have been all too few therapeutic advances in recent decades, something he blames in part on the “perverse” role of patents.
“When you do patents early, sometimes you have a paradoxical effect of slowing down discovery and development,” he explains. “We see open science as a means to expand the impact of our research by sharing it with the global community.”
In 2015 the MNI undertook an 18-month consultation with its scientific community to determine the parameters of this openness, but a primary feature would be the absence of intellectual property. The result was dubbed the Tanenbaum Open Science Institute, named after the family that donated $20 million to kickstart what amounts to a very ambitious social experiment.
“Open science isn’t free science,” Rouleau points out. “It still costs money.”
That money goes into building the enabling platforms to allow open science practice — neuro-informatics for efficient data-sharing, biobanking, open drug discovery, open publishing — something that is far from trivial in an era where consumers still struggle with the vagaries of choosing an acceptable Web browser. A team led by McGill University law professor Richard Gold is developing elaborate tool kits to help its intended users determine what they want from an open science platform, and more importantly, how they will measure the impact that such a platform has on their work.
“We’ve gone out into the community to find out what’s needed — what people hope that open science produces, or perhaps fear that it produces,” says Gold, who also chairs a committee that is evaluating the success of the MNI initiative. “We want to identify where partners are and then look at the results they produce. Then in one or two years if we get enough of these institutions doing this we can start seeing trends. What we really want to do is create a set of data so that in five years someone could look back and ask, ‘what did open science produce?’”
Even if the answer to that question turns out to be disappointing, Gold insists that reviewing open science in this clear and dedicated fashion will yield useful findings. If nothing else, it should shed some light on how to draw the line between a productive exchange of scientific ideas and the patent protection designed to ensure that commercial enterprises can afford to turn those ideas into products, such as drugs that benefit society on a broader basis.
Those two objectives need not be mutually exclusive. The SGC has created its own spin-off company, Meds4Kids (M4K) Pharma, which is seeking cures for childhood diseases that pharmaceutical manufacturers regard as too rare or risky to be worthy of making their own research investments. M4K will not be seeking any patents on its work, which it is encouraging outside researchers to share, but it will claim rights to the resulting data for the approval of any successful therapy. The profits from those products are then to M4K’s parent firm, Agora Open Science Trust, which will use any proceeds to further open science and the public good.
“Data protection is the package the drug sponsor files with the FDA or Health Canada, to show that the molecule or biologic is safe and efficacious,” says Gold. “Nobody’s allowed to duplicate that file, but anyone can take the same molecule and re-do the clinical trials. You’re not limiting people doing what they want with the information.”
Linux, the popular open-source computer operating system, is protected in much the same way. While its core code remains free of any intellectual property, the many different ways of applying that code have been secured with “defensive” patents. Anyone trying to enforce a patent on Linux, such as suing users of the operating system, would instead find themselves being sued by one of these patent-holders, who have a legitimate claim against anyone trying to assert that kind of control over the core code.
“We’re getting more sophisticated in our thinking about how to protect that freedom to operate and people’s real ability to use the data freely,” says Gold. “It goes well beyond the health sciences, into areas like agriculture and environment.”
End users, regardless of how they might feel about the disposition of patent rights, may well enjoy the ability to dig deeper into their areas of interest.
“So often you’re trying to figure out in a paper where are the data that were used, and you can’t find it or you have to extrapolate from given figures,” says Elizabeth Edwards, a professor in U of T’s Department of Chemical Engineering and Applied Chemistry. “So I like the move to curated data repositories.”
Such open archives will not resolve all the difficulties associated with making data more widely available, according to Martha Crago, McGill’s Vice-Principal of Research and Innovation.
“There’s going to be variations among different disciplines, there are going to be particulars within a discipline,” she says, recalling the questions raised by a linguistics researcher whose work was premised on recordings she had made of her daughter’s earliest utterances during the first few years of her life. Those recordings are now part of the official research record, something the now-adult daughter — who has spoken on the matter at academic conferences — points out that she never approved.
That may seem like a milder ethical breach than He Jiankui’s spawning of genetically modified children, but Crago notes that it falls on the extensive spectrum of behaviour that is still being defined to cover what is meant by “open” science.
“There’s ethical concerns and real data storage concerns,” she concludes. “Then there’s all kinds of interesting conundrums that it can lead to — or beauties that it can lead to.”