Few things have greater potential to influence access to research resources than patents and the policies that determine when they are awarded. Patents are intended to encourage innovation by guaranteeing that innovators are appropriately compensated for their work. Patents are also intended to encourage openness. Without patent protection, inventions and discoveries with commercial potential would often be kept as trade secrets. The patent system offers the inventor legal ownership of an innovation in exchange for putting a description of it into the public record.
Patenting an invention keeps people from using it without paying for it, or at least asking permission, but it does not mean that the invention cannot be used at all. “There is a big difference between being available free and being available,” Craig Venter of the Celera Corporation noted. Indeed, if an invention demands expensive development before it is useful to anyone, patenting the invention might be the only way it ever becomes accessible.
Nonetheless, patenting can slow or stop access to some innovations, particularly basic discoveries and inventions that are of value to researchers on the leading edges of their fields, so scientists are especially sensitive to patent policy regarding this sort of fundamental work. Perhaps the best example is a debate that has been roiling the molecular biology research community for some 7 years. The resolution of the debate will come not from the scientific community, but from the US Patent and Trademark Office (PTO), patent attorneys, courts, and perhaps Congress. However, it will be scientists who are most affected by whatever decision is reached. And not just molecular
Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 12
FINDING THE PATH: Issues of Access to Research Resources 2 Patents Few things have greater potential to influence access to research resources than patents and the policies that determine when they are awarded. Patents are intended to encourage innovation by guaranteeing that innovators are appropriately compensated for their work. Patents are also intended to encourage openness. Without patent protection, inventions and discoveries with commercial potential would often be kept as trade secrets. The patent system offers the inventor legal ownership of an innovation in exchange for putting a description of it into the public record. Patenting an invention keeps people from using it without paying for it, or at least asking permission, but it does not mean that the invention cannot be used at all. “There is a big difference between being available free and being available,” Craig Venter of the Celera Corporation noted. Indeed, if an invention demands expensive development before it is useful to anyone, patenting the invention might be the only way it ever becomes accessible. Nonetheless, patenting can slow or stop access to some innovations, particularly basic discoveries and inventions that are of value to researchers on the leading edges of their fields, so scientists are especially sensitive to patent policy regarding this sort of fundamental work. Perhaps the best example is a debate that has been roiling the molecular biology research community for some 7 years. The resolution of the debate will come not from the scientific community, but from the US Patent and Trademark Office (PTO), patent attorneys, courts, and perhaps Congress. However, it will be scientists who are most affected by whatever decision is reached. And not just molecular
OCR for page 12
FINDING THE PATH: Issues of Access to Research Resources biologists; the decision of where to set the threshold for patenting could have repercussions for other fields, such as synthetic chemistry. The debate commenced in 1992 when PTO rejected a patent application from NIH for thousands of expressed-sequence tags (ESTs). ESTs are short stretches of DNA that can be isolated rapidly and in great number and used to identify genes. They are, in effect, gene fragments that are long enough to be characteristic of particular genes but short enough to be manipulated and cloned easily with the standard tools available to molecular biologists. PTO refused to approve the EST patent application on the grounds that the function of the DNA was not known. But the matter did not end there. Private genome companies have since filed patent applications for hundreds of thousands of ESTs. One goal is simply to establish priority, to establish that a particular company was the first to isolate a particular gene, even if only a very small part of it, and even if the company had no idea what gene it was when it filed the application. But another goal is to gain patents and thus establish an interest in whatever else is done later with the genes identified by the ESTs. More recently, companies have been rushing to patent single-nucleotide polymorphisms. Differences in the human population are produced by variation in the nucleotide sequence of their DNA, which is composed of four types of bases—adenine, guanine, cytosine, and thymine—A, G, C, and T. If the DNA sequences of two individuals are compared, or if the maternally and paternally derived chromosome pairs of one individual are compared to each other, there will be differences. If at a particular position on the DNA sequence the maternal chromosome contains a ‘G' while the paternal chromosome contains a ‘T', each surrounded by otherwise identical sequence, that difference is called a single nucleotide polymorphism—a SNP. SNPs (pronounced “snips”) can be found on average every few hundred nucleotides across the entire human genome (of a few billion nucleotides) so there are literally millions of SNPs that can collectively distinguish each of the 42 human chromosomes carried by different individuals. Those differences in sum produce a fingerprint, a ‘SNP Map', which can be used to trace patterns of inheritance of disease predisposition genes that contain, or are closely linked to, particular SNPs. It is the potential for mapping complex characteristics by variation among individuals in their genome-wide SNP maps that is engendering considerable excitement, and commensurate concern about access to this powerful research material. Although much effort will be needed in the future to understand the role of SNPs in relation to disease, SNPs, like ESTs, can be identified now with relative ease using the proper tools, with little effort beyond the initial set-up. It is precisely the latter characteristic that raises questions about whether ESTs and SNPs should be patentable, said Steve Holtzman. The power of modern molecular biology has automated this particular type of discovery to the point where a laboratory with the right equipment can, for example, uncover thousands of ESTs a day. Traditionally, however, a patent has been issued for
OCR for page 12
FINDING THE PATH: Issues of Access to Research Resources something that demanded creative and original work; there, Holtzman noted, is the rub. “Do you feel a cognitive dissonance there—‘automated discovery and invention'? What does it mean to have 10,000 creative, original, useful ideas a day? I think there's a problem here, and it has to do with conceptual impossibility.” Certainly, the originators of our patent laws never conceived that they might be applied in this way, but the patent system has evolved in such a way that granting patents to such automated discoveries is almost inevitable. Venter, Holtzman said, “came up with the brilliant idea of ESTs. I submit that that was the invention, the method.” The individual ESTs churned out by the thousands are not the invention, he said. “But that's not consistent with patent law. Patent law says that the composition of matter is the embodiment of the idea, and you get the grant on the embodiment of the idea, on the composition of matter. Furthermore, case law has made it very clear—and this is intuitively true—that the method of invention doesn't compromise the invention.” In other words, it does not—and it should not—matter whether you worked for years to identify a single EST or pulled it out of a machine with thousands of others. But if such automated invention is patentable, it will be possible for a few companies with a lot of money and a lot of machines to lay claim to huge areas of knowledge before these areas are ever explored more than cursorily. It is not just ESTs and SNPs that are in play, Venter argued, but whole genes as well. Nor does the problem stop with molecular biology. There are perhaps 100,000 small molecules —with various combinations of carbon, hydrogen, oxygen, nitrogen, and a few other atoms—that could be useful as drugs, and combinatorial chemists are rapidly—and automatedly—trying out all those possible combinations. “People are getting patents granted daily right now that claim thousands and thousands and thousands of those molecules.” This sets the stage for so-called submarine patents—patents granted on the composition of ESTs, SNPs, genes, or small molecules—that will one day surface to exert their ownership rights when discoveries are made by others about the function of the these biological and chemical entities. Not only does that seem unfair and contrary to the original intent of the patent system, but it also makes researchers and companies less willing to invest the effort in pursuing discoveries that could be waylaid by submarine claims. So far the PTO has awarded one EST patent. It is a broad patent that claims rights to any gene that contains the EST, even though the gene is not yet known. Venter argued that the patent is probably not worth the paper it is printed on, because “the patent had very low accuracy data and lousy informatics and claimed things that don't really relate to the sequence.” Holtzman countered, “It might be not worth the paper it's written on, but it will cost you several hundred thousand dollars to litigate. So it's worth something.” It is, if nothing else, a disincentive for anyone besides the patent holder to pursue the gene that contains the EST.
OCR for page 12
FINDING THE PATH: Issues of Access to Research Resources Thomas Caskey agreed. “I can tell you that this is a major issue right now, because broad claims are being filed. The patents are being awarded; therefore, using that gene in the commercial world for any type of utility gets greatly complicated because the corporation lawyers will tell you that there is great jeopardy in proceeding with the use of that gene if you do not have clearance.” Nonetheless, under existing patent law, the ESTs and similar “discoveries” are patentable. The most relevant issue affecting their patentability is their utility—PTO demands that any invention or discoveries have some potential use if it is to be patented. But even if the biologic function of a gene, SNP, or small molecule is unknown, it can still have some utility. Consider the SNP, for example. Celera is planning to sequence five human genomes, Holtzman noted. “By definition, they will have 80% of the SNPs with a 20% prevalence in the population. Do they have utility? Yes. They are mapping reagents.” That is, they offer a way to distinguish one person's DNA from another's, which can be important, for instance, in tracking down the genes that cause a disease. “Any utility is sufficient for a patentability claim, said Holtzman, adding sarcastically, “All of us have thousands of SNPs. We could file on them tomorrow. ” The situation for small molecules is similar. When researchers look for drugs that will attach to a particular protein in the body, they often start with one drug that works to some degree and comb through others with similar structures, looking for the one with the best performance. Thus, it is useful to have libraries of small molecules available for searching. “They have utility,” Holtzman said. “Someone will buy those libraries.” The best way to proceed, Caskey suggested, might be to finetune how PTO awards patents. “We're not going to reverse patent law, so my simplest solution to this would be to ratchet up the specificity and the demonstration of the specificity of utility,” that is, require the patent applicant to demonstrate explicitly the utility of the invention and then grant the patent for that utility alone. If the only known use for a particular small molecule is as part of a library of similar small molecules, grant a patent for that use; if someone later discovers that the molecule is a useful drug for schizophrenia, allow another patent for that specific use. Then, Caskey said, “as you move down the pathway, there is an opportunity for protection with new discovery to be able to go forward to the utility.” Holtzman did not think that remedy would work. “In an ideal world,” he allowed, “the grants of the composition-of-matter patent would only go along with ‘real utility.'” But, he said, he has little faith in the ability of PTO, Congress, or other institutions to get it right. “PTO was operating that way and basically saying with respect to drug molecules, ‘Do a Phase III study and get it registered by the Food and Drug Administration, and then you will have shown its utility.' It backed off from that 2 years ago.” It had set the bar so high for a patent that it was creating more problems than it had solved.
OCR for page 12
FINDING THE PATH: Issues of Access to Research Resources A different approach that many have suggested for the genome field would be for the federal government and other entities to pay for putting as many data into the public domain as possible, thereby making the data unavailable for patenting. In 1994, for instance, spurred by worries that much of the human genome sequence data would be patented, Merck set up a project to sequence human DNA and deposit it in Genbank, the public database of genes run by the National Library of Medicine. And the NIH National Center for Human Genome Research encourages the researchers who receive its grants to deposit their DNA sequences into a public database as quickly as possible and not to seek patents on the sequence information itself. That does not mean, however, that discoveries made about the meaning and usefulness of raw genetic data would not be patentable. “We don 't think by publishing the human chromosome sequence itself we're blocking others from making the kind of key discoveries that have been talked about,” Venter said. “My understanding is that publishing the human chromosome sequence itself will have no impact on cDNA or protein patents.” For example, Venter has filed for a patent on a human gene that codes for a new serotonin receptor; he found that gene by searching through a human chromosome sequence that had been deposited in Genbank. “The best patent-attorney advice we can get is that those should be valid claims.” Furthermore, Caskey noted, each step in the process of moving from a gene or protein to a commercial product offers its own opportunities for patent protection. “If we want to try to keep the roadway open for discovery,” he said, the information that underlies everything else—the sequence data—should be in the public domain, and patents should be reserved for discoveries and inventions that go beyond that basic information. Most recently, Celera, Venter's company, has announced plans to determine all the 3 billion base-pairs of the human genome and put that sequence information on the Internet, making it freely available to anyone who wants it. If all goes as hoped, Celera will have the entire human genome finished by the end of 2001. As a warm-up, the company plans to sequence the smaller genome of Drosophila melanogaster, the fly used by many biologists as a model organism, and make that sequence information available by the end of 1999. Once the entire human genome—and several other genomes as well—is available, those discoveries and innovations should explode, and Celera plans to profit by serving prospectors who are looking to mine the various genomes. Although it will place no restrictions on the use of its data—anyone can download unlimited human sequence information free, freely distribute it, and never pay royalties on any inventions stemming from it—Celera will have developed a resource that no one is likely to duplicate. “If somebody were going to try to duplicate the data center that we're building, it would cost around $60 million just to build the housing for the hardware, and then they would have to pay for data generation and support.” Thus, people who want to work with the DNA data will come to Celera, Venter said.
OCR for page 12
FINDING THE PATH: Issues of Access to Research Resources Venter views the commercial strategy of on-line information providers like Lexis-Nexis, (which maintains large databases of government publications, legal case histories, and more), as the model for Celera. Once such a vast amount of data is compiled, users will want convenient access to this comprehensive body of information, and be willing to pay for it. “Pharmaceutical companies are paying pharmaceutical prices for unique early access,” he said. “That's at one extreme. The other extreme is that the data will be on our Web site and freely available to academic researchers. In between are the people who want the added value of all the comparative data and all the other information; we're thinking of subscription prices of around $5,000-20,000 a year for research laboratories, which would be compatible with what academic scientists are paying for other research tools and software systems.” Not everything Celera produces will be available, however. In particular, although the company does not believe that sequence data should be patented, it is taking a wait-and-see approach to its SNPs database. “We're eager to see what happens with intellectual-property protection on polymorphic variations,” Venter said. “I think they're very important for screening and for a wide variety of tools. What's driving it, obviously, is the pharmaceutical industry, which wants to save billions of dollars off the cost of developing drugs.” With billions of dollars in play, Celera will, at least initially, keep its SNPs as trade secrets, Venter said. “But because the basic data are accessible to subscribers at reasonably competitive rates, we think that they will actually be broadly available to a wide array of scientists.” That, to Venter, is the bottom line—not whether something is patented, but whether it is accessible. In discussing patents, he said, many people forget that the patent system makes possible commercial development of many of the research resources that scientists depend on. “If you look at science in this country versus in, for example, the former Soviet Union, the biggest difference is that we have tremendous industry support for what we 're doing”—and this industry support is available because the patent system guarantees that companies can profit from developing such things as restriction enzymes and other tools that researchers use, as well as drugs and other products. “If composition-of-matter patents are denied on DNA, it would certainly affect the Amgens of the world and the Genentechs that make incredibly important drugs that have saved millions of lives in this country.” Instead of getting hung up in an emotional debate about whether it is right for someone to “own” our genes, Venter said, the better approach is to ask how well the patent system is working to promote access to scientific discoveries. There might not be agreement on the answer, but there should at least be agreement on the question.
OCR for page 12
FINDING THE PATH: Issues of Access to Research Resources BOX 2 THE HIGH COST OF HIGH TECHNOLOGY In addition to intellectual-property issues, a major obstacle to research resources for many scientists is cost. Research tools can be divided roughly into three groups, observed Michael Snyder. “ There are cheap technologies, such as restriction enzymes and the polymerase chain reaction (PCR), that spread throughout the community like wildfire. They get invented, and even if you try to stop their unlicensed use, you can't. “On the other end of the spectrum are technologies, such as linear accelerators and space stations, that are so expensive that we can afford to build only one or a few of them, and everyone must share. “And then there are the technologies in between, such as confocal microscopes, automatic sequencers, and micro-array technology. These are fairly expensive. Individual laboratories—certainly individual academic laboratories—can't go out and buy these easily.” The technologies in the middle group are extremely powerful, allowing researchers to do things that would otherwise be impossible, so it is important to make them as widely accessible as possible. Yet, Snyder said, “some of these technologies aren't being distributed well.” For instance, he looked at 28 recent papers by US scientists describing research that used micro-array technology. “Of those 28, 24 came from very large research centers, companies, or a very well funded laboratory. Only four came from what I saw as academic laboratories. So we have a situation of haves and have-nots. A few laboratories are using the technology or can collaborate with people who can do it, and there are a lot of people who want access but cannot get it.” For such situations, Snyder said, it would make sense to set up “minicenters” around the country that make a technology accessible to a much broader range of researchers. “Twenty small centers could blanket the country,” he suggested. “I don't think it would be very expensive to set up this particular kind of technology. For example, $200,000 would certainly cover the cost of one center for micro-array technology, and with matching funds, the center could be even more productive. If you had 20 centers scattered in various geographic locations, you could get all this technology out there for relatively modest cost. Four million dollars is pretty modest when you think about the impact that this technology has on things and how much it cost to invent it in the first place.” The same argument applies to a number of research tools, Snyder said. “If you think of the nature of science now, a lot of these technologies are coming out of big laboratories or big centers, which devise expensive technologies that individual investigators can't afford.” To make the most of these technologies, “we'll need groups of people to have access, and I suggest that minicenters would be wonderful avenues for dispersing useful information and technology.”