Open Source Biotechnology
A New Way to Manage Scientific Intellectual Property
by Janet Hope

Since the 1980’s, the life sciences have undergone a process of rapid commercialization. The legal mechanism for this process of commercialization has been the expansion of intellectual property (IP) protection to inventions that were previously regarded as unpatentable. The result has been a literally exponential increase in applications for biotechnology patents.

These patents not only protect inventions that are valuable as end products; they also protect early stage inventions and research tools. Advances in biotechnology require the use of many of the latter, for which researchers must obtain licenses from patent owners. A good example is “golden rice”, which utilized more than 70 different patented procedures and processes. To get permission to use all of these tools, scientists enter into multiple negotiations for each piece of IP. These mounting transaction costs can retard, and in some cases completely undermine, their scientific projects. Even if they are not prevented from pursuing research itself, institutions may find that the rights of other IP holders prevent them from commercializing the fruits of their labor.

In biomedicine, there are considerable social costs associated with working within this expensive proprietary system. These stem from the fact that such costs are beyond the resources of the smallest participants, or would-be participants, in the industry. Market forces will naturally tend to direct efforts by big private sector players to where there is the most substantial return on investment. This means research goals are inevitably being narrowed to those that will be most profitable, though not necessarily most useful. Thus, it is often not commercially worthwhile for the biomedical industry to devote significant resources to addressing medical or social needs, such as drugs for very common diseases like tuberculosis or malaria.

Similarly, in agriculture, breeding strategies will be oriented towards major crops in developed country markets, not towards finding genetic traits with characteristics that are useful to poor farmers. The last few years have also seen a series of mergers and acquisitions that have dramatically consolidated the industry, with a huge portion of fundamental research tools ending up in the hands of a tiny number of big multinationals. This level of industry concentration has inevitably led to the overpricing of technologies and the exclusion of innovative start-ups and public sector institutions. This, in turn, means that smaller firms can’t get a foot in the door.

This situation has been described as a “tragedy of the anticommons.” In contrast to the tragedy of the commons, when a public resource is overused because there is no one owner to regulate it, a tragedy of the anticommons occurs when a resource is underused because it has been divided up by a number of owners who may not be willing to agree or cooperate with one another.

Open Source Software

The costs associated with proliferating IP rights threaten to exclude all but a handful of firms from shaping the direction of technological change. To anyone who has followed the fortunes of the software industry over the past decade, this statement will have a very familiar ring. The difference is that, in the software industry, an alternative way of dealing with IP has emerged, one which has been very successful in some areas at preventing the kind of industry consolidation that has occurred in agricultural biotechnology . That approach to IP is known as “open source,” and an analogous approach might offer a partial solution to these kinds of problems in the biotechnology industry.

There are two distinct aspects to the concept of open source. The first is that open source constitutes a specific, relatively well-defined approach to the licensing of intellectual property. Second, open source represents a characteristic methodology for developing technology.

To understand open source licensing we need to understand the style of copyright licensing known as “copyleft.” The most commonly used copyleft license is the General Public License (GPL). Under the terms of the GPL, the copyright owner grants users the right to copy, modify and distribute the licensed software without having to pay a fee to the owner. The only proviso is that users must then license any modified versions they create under the same terms as the original.

By using a copyleft license, one can make the property freely available and still retain some control over it. In the case of the GPL, that control is used to ensure that even as the licensed software evolves into forms that might be sufficiently new to attract copyright protection in their own right, the software remains available for everyone to use or build on.

The idea of open source represents a further evolution of the free software concept. To be open source, software must be distributed under a license that guarantees users three rights without charging any fee in return.

First is the right to have access to the source code—that is, to have access to a version of the program in a form that can be understood and modified by a user. Second, users must have the right to make as many copies of the program as they like and sell them or give them away without having to pay for doing so. Finally, the license must allow the creation of modifications and derived works and must allow the user to redistribute these new versions under the same terms guaranteed by the original license.

The program that is usually regarded as the archetypal open source software is GNU/Linux, created by Linus Torvalds. Although the Linux kernel is probably the most prominent and successful, it is only one of thousands of open source projects now being used in the software context.

Typically, an open source project starts when an IP owner makes protected technology widely available under an open source license. Users then have access to a toolkit that can be used to extend and improve the technology and apply it in a range of ways at their discretion. They are free to sell any products and services they produce using the technology, but may be required by the terms of the license to contribute any improvements they make to the technology itself back into the common pool. This ensures that the modified technology can be used by other people on the same terms as the original open source license. In this way, the technology is not kept in the control of a single individual or company, but instead can be developed, produced, distributed and supported by a community of users, each of whom has the capacity to shape the direction of development in accordance with their needs. Development is surprisingly rapid because the IP is dealt with in a way that facilitates access and improvements to the technology. Thus, open source is a way of allowing more people to influence the pace and direction of technological development.

Can Open Source Licensing Work With Biotechnology?

When I spoke to Bruce Perens, who helped define the basis for open source development in his aptly titled document, The Open Source Definition, he took the view that the open source biotechnology movement does not aim to create a particular legal framework. Instead, it is a form of social engineering. There is no question that one could produce a legally binding open source license in biotechnology if one wanted to—the real question is whether anyone will use it.

The different proprietary regimes that prevail in the software and biotechnology contexts are important to consider in answering this question. Both software code and biotechnology innovations are protected under a mixture of licensing systems , but the primary one in software is copyright, whereas in biotechnology it is patents. The cost of patent protection can be substantial, whereas copyright protection arises automatically and without cost to the owner. Also, patent fees are usually at least partly recovered from licensees under the remuneration clauses in a proprietary license.

Second, standardized licenses appear to be important for keeping transaction costs low in open source software, but this approach may be less applicable outside a digital context. Biotechnology innovations are far more diverse in terms of composition than software, which is essentially non-physical and instantly reproducible. Defining rights in living biological materials, given their capacity for self-replication and mutation, is difficult. Determining what constitutes an improvement to a licensed biological technology is also challenging. This aspect would be especially critical in open source applications. As stated earlier, open source licenses generally require that improvements to the technology be made available to the other users. Naturally, this is far more difficult when the medium is biological matter, as opposed to digital information.

To explore how open source might translate into the biotechnology context, it is necessary to characterize it in terms of generalized principles, as distinct from software-specific features. Although it is becoming a popular subject of study for people in many disciplines, no unifying principle has yet emerged as the dominant approach. I have chosen to view open source development through the lens of a relatively new theory from the field of innovation management, known as “user innovation” theory.

The seminal work in the user innovation literature was Eric von Hippel’s 1988 book, Sources of Innovation. However, the literature also draws on earlier empirical studies in the broader field of innovation studies. There are five key elements to this theory: user innovation, free revealing, collective invention, distributed production, and community support.

User Innovation

The user innovation literature draws a distinction between users and manufacturers of an innovation, the categories being defined according to the type of benefit an innovator expects to attain from his or her work. Users primarily benefit from using the innovation, while manufacturers primarily benefit by selling it. The conventional assumption is that manufacturers, rather than users, are likely to be the main innovators in any given field, for the simple reason that “making one and selling many” items is assumed to be the most profitable way to exploit a piece of technology. By contrast, the user innovation literature points to empirical evidence that users, rather than manufacturers, are in fact the primary innovators in many contexts, including open source software.

In such contexts, the information required to come up with new technological developments is “sticky”—it is costly to transfer from one person to another. It makes more sense for a user who already has most of that information to invent a new piece of technology than it does for a manufacturer, who would have to invest in researching what users need and how that particular innovation would function in a given industrial setting.

Biotechnology research and development (R&D) is an area where it often makes sense for users to do most of the innovating. This is because biological information is particularly sticky; it is complex and highly location-specific. This is partly a function of the nature of biological materials and partly a function of the early stage of development of many biotechnology inventions.

Free revealing

The second assumption is that any “free revealing” or uncompensated spillover of proprietary knowledge developed through private investment will reduce the innovator’s profits from that investment. It is therefore assumed that innovators will avoid such spillovers as much as possible. The user innovation literature challenges this assumption. Although the proprietary approach is often the most profitable way to exploit an innovation, user innovation theory shows that this is not always the case.

There are disadvantages to the proprietary approach. Secrecy is one example. Proprietary knowledge must be kept secret for it to remain that way, but this only makes sense for inventions that cannot easily be reverse-engineered. Even though there is no limit to the term of trade secret protections in principle, in practice, most development secrets can be compromised, sometimes very quickly. The costs of preserving secrecy can be significant and must be balanced against the benefits of exclusive access. Further, when it comes to licensing an innovation, a dilemma arises. IP owners want to sell as many licenses as possible to generate revenue, but if the information is disclosed to too many users, the benefit of trade secrecy protection may be lost.

There are also several ways in which a self-interested innovator can benefit from a free revealing approach. First, free revealing of IP may establish a non-proprietary technology that encourages the purchase of proprietary technology with which it works. To use a software analogy, a company may make an operating system free to users because it intends to sell other software that is exclusively compatible with this operating system. These secondary pieces of proprietary technology can offset any monetary losses from the non-proprietary technology.

Free revealing may also facilitate improvements to the core technology. If the original innovator then gains access to those improvements, this represents a cost saving in R&D for that company.

In addition, by revealing its IP, a company may generate a favorable reputation that is useful in selling associated offerings, by enhancing brand value or improving the company’s ability to attract and keep high quality employees.

It is often assumed in discussions of open source that the benefits of adopting a free revealing strategy for exploiting innovations must be unique to software or other information goods, but such benefits can apply to physical goods as well. Biotechnology can benefit from free revealing, and some companies are already beginning to implement such practices to this end.

As mentioned earlier, one can use non-proprietary products as enticements for other proprietary goods or services. Companies are often willing to contribute to the creation of open databases in order to attract customers to a host Web site that offers additional commercial content. Others give away the rights to use valuable cell lines and experimental animals, and sell consulting services on how to maintain them. Microarrays, a way for conducting many experiments at once on a DNA or protein “chip,” are another potential field, as proprietary chips are too expensive for some institutions and applications. Developing open standards for manufacturing microarrays could produce a proprietary market for tools that would conform to those standards.

Another way is to pre-empt the establishment of a proprietary standard. The most obvious example of this strategy is the participation of companies in the SNP consortium, which pays academic scientists to place genome sequence data in the public domain. For these companies, giving away data is not a charitable act—it avoids having to negotiate IP access among themselves and with other companies down the line. Interestingly, the human genome sequencing project considered adopting open licenses, but the idea was abandoned because it was decided that any restrictions on the data, even in the form of a license designed to ensure it stayed non-proprietary, would create a dangerous precedent. In that case, open source development was successfully taken to its extreme.

An even more ambitious proposal for large-scale collaboration has been made by Steve Maurer, an economist at UC Berkeley. Maurer suggests harnessing and combining volunteer efforts across the life sciences community to research potential malaria drugs that would then be made publicly available. Maurer argues that an open source approach would reduce the total costs of drug development. As highly trained volunteer labor would perform the research, sponsors could avoid overpaying R&D costs, which are difficult to estimate in early stages. In addition, because the IP would be available to everyone, any company could manufacture the drug, and the resulting competition would keep down the market price for the completed product.

Biotechnology companies can also enhance their reputations by using open source techniques. One example of this strategy is the contract research company Millennium Research. Millennium makes much of the technology it develops available to the general public, boosting its reputation for innovation and expertise, as well as for user-friendliness and social-mindedness. This kind of attitude seems to resonate with users.

Along with the benefits of adopting a free revealing, non-proprietary strategy for exploiting an innovation, there are costs to contend with as well. Opportunity costs are the gains that an innovator could have made by adopting an exclusive proprietary approach. Actual costs include the expense of diffusing an innovation. Clearly, choosing the best strategy for exploiting any particular innovation involves weighing the costs and benefits of a proprietary versus a non-proprietary approach. It is a trade-off, and in many cases, in biotechnology as elsewhere, the balance may tip in favor of the traditional proprietary approach. This is generally the case in the pharmaceutical industry.

Normally, the opportunity cost of adopting a free revealing approach is not prohibitively high, as patents have a limited term and can be easy for other industry members to circumvent. However, the pharmaceutical industry has structured itself so as to make patent ownership particularly profitable. It has successfully pushed for patents that are broad enough to effectively cover not just a particular molecule that has value as a drug, but all the variations of that molecule that might be effective. This means that pharmaceutical patents are almost impossible to invent around. Combined with legal tactics for extending patent terms, the opportunity costs of giving up an exclusive proprietary approach to drugs in favor of an open source approach are likely to be too high for big pharma to be interested. As stated earlier, cases like these illustrate some of the negative effects of proprietary IP regimes.

Collective Invention

Collective invention comes about when enough innovators adopt free revealing approaches to producing a particular innovation. The result is a cycle of free revealing developments which may or may not be reinforced by a copyleft-style licensing regime.

There are several conditions that favor the development of a collective invention regime in any given industry; two are particularly relevant to biotechnology. First, collective invention is more likely to take root where R&D is expensive and its outcomes are uncertain. This is because each firm’s expected payoff from working privately may not be enough to cover the costs, or it may be too expensive to do at all. In general, biotechnology is significantly more capital intensive than other fields, and often the only way to get a desired end product is to share the burden.

Collective invention is also more likely to take place where standardization within an industry makes collective learning easier. While genetic engineering is inherently more complex and diversified than software engineering, there are certainly circumstances to which open source development can be applied.

Distributed Production and Community Support

The final two elements of an open source development approach, according to the user innovation literature, are distributed production and community support. Both are possible in a biotechnology context.

Community support for collective invention is the crux of open source development. As stated earlier, there is no question that one could implement the legal framework for open source development in the life sciences. The real question is whether anyone would be willing to use it. Many of the community-based incentives and support structures that help to drive open source software development have an equivalent in the life sciences. While the two fields are markedly different, the advantages displayed in the former can definitely entice a community of adopters in the latter.

The potential for distributed production is less clear-cut. Distributed production is a non-issue for information-based products, as it can be transported and reproduced by users with no real cost. For physical goods, production and distribution involve economies of scale that are best exploited by manufacturers. This certainly seems to apply in some areas of biotech, such as the large-scale manufacture of pharmaceuticals, a field heavily entrenched in the principle of proprietary IP. On the other hand, it is clearly not a major factor in relation to self-replicating biological materials, such as seeds. In fact, the inherently dispersed and localized nature of the agricultural enterprise suggests that a distributed approach to technology development may make much more sense than a centralized approach.

Conclusion

Proprietary approaches to intellectual property in biotechnology have the effect of consolidating knowledge and research tools into a smaller number of hands, stifling the diversity of research and increasing costs to users. Open source approaches to technology licensing and development offer the possibility of a partial solution to these problems. Open source approaches won’t always be applicable to material media, but there are many areas where they could effectively work.

The biggest determinant of open source’s adoption is whether the balance of costs and benefits will make it more attractive to IP owners than the proprietary approach. Of course, this is less likely the more a manufacturer is entrenched in proprietary IP. Thus, it is important to note that even if there are only one or two adopters of open source in a given industry sector, the actions of those adopters can have a big impact. By undermining customers’ willingness to pay for access to tools from a proprietary source which they can get at a lower cost from an open one, a small number of open source adopters can shift the balance of competition in a sector away from proprietary technologies. In this sense, open source principles have the power to transform industries. This is why Microsoft is wary of Linux, and it is also why open source has the potential to break the IP logjam in biotechnology.

People often conflate the successes and failures of open source software with the possibility of open source biology. In this connection, people bring up the current copyright litigation surrounding Linux distributors and the overall non-profitability of such companies as illustrations that the future of open source software is uncertain. While it is true that there are not yet any stand-alone open source software businesses that are actually profitable, this is also largely the case for the traditional business model in biotech. Despite the hype and the high market valuations of the late 1990’s, biotechnology has not yet been a profitable industry overall. In any case, it is not necessary to see open source as a stand-alone business model. Instead, it should be viewed as a business strategy that may be used to complement other strategies.

In the end, the proof for the viability open source biotechnology is not tied to the ultimate success of open source software. Open source software is simply the basis for an analogy—the seed of an idea rather than a rigid formula for success.

Janet Hope graduated with first class honours degrees in law and biochemistry/molecular biology from the Australian National University in 1995 and was admitted to practise as a Barrister and Solicitor of the High Court of Australia in January 1997. She is currently completing her doctoral thesis on Open Source Biotechnology and has recently won an Australian Research Council grant to pursue post-doctoral research on co-operative intellectual property management