INTELLECTUAL PROPERTY: Patents, Secrecy, and DNA Robert Mullan Cook-Deegan and Stephen J. McCormack

Robert Cook-Deegan and Stephen McCormack

Policies governing DNA sequence data access span a wide spectrum. At one extreme, private firms generate data that are used internally or licensed to pharmaceutical firms for multimillion-dollar annual fees. These data are effectively unavailable to most potential users, kept private through licensing terms and nondisclosure agreements. They will achieve social good in the form of products eventually. Their contribution to useful knowledge through academic research, however, is reduced by the high financial barriers to access. For Incyte and Human Genome Sciences and several other firms, the main subscribers are pharmaceutical and biotechnology firms with large R&D programs that can afford the databases access fees. Both firms do have academic collaborations, but they are limited and relatively infrequent (and terms of the agreements not publicly available). Celera's subscription strategy includes pharmaceutical and biotechnology firms, but also more academic and nonprofit groups that agree to pay fees and abide by terms limiting disclosure to third paries. Celera offers the possibility of direct consumer subscriptions in the future.

At the other extreme, government- and nonprofit-funded high throughput sequencing groups release sequence data daily, based on a 1996 agreement reached in Bermuda. The major centers involved in the "public genome project" agreed to abide by the Bermuda Accord, and this includes the world's largest public-domain sequencing operations. It covers only a few dozen grantees and contractors, however, among the thousands of molecular biology research centers throughout the world that do DNA sequencing in the course of their research.

The Bermuda Accord is beneficial because without it, the wait for information sufficient to meet patent criteria from high throughput sequencing programs would lead to long delays, and thus be a serious drag on science, undermining the publicly funded sequencing programs' very purpose. One inevitable consequence of the Bermuda Accord is, however, an asymmetry between those who follow its precepts and those who do not. Those using public data but not contributing to it gain an advantage--and the larger the sequencing capacity and the greater the bioinformatic competence of the private user, the greater the advantage. This asymmetry is a root cause of acrimony between Celera and the public genomic sequencing efforts, but it follows directly from policies the public project participants rightly imposed on themselves. The solution is not to change terms of the Bermuda Accord but rather for all users to acknowledge the considerable value of public domain information and to welcome, rather than cavil at, private sector use.

Intermediate policies are emerging, such as the recent agreement to release data on protein structure after patent application or 6 months, whichever occurs first.

The National Cancer Institute's Cancer Genome Anatomy Program (CGAP) includes a component devoted to sequencing full-length genes. CGAP is the largest public domain effort comparable to gene-sequencing programs at Human Genome Sciences, Incyte and other private companies. CGAP is administered as a set of contracts. NCI obtained a declaration of "exceptional circumstances" under the Bayh-Dole Act, so that contractors assign patent rights to NCI, and rapid publication of sequence data is a contract condition ("Determination of Exceptional Circumstances Under 35 USC 202(a)(2) and 37 CFR 401.3(a)(2) and (e) for the Full-Length cDNA Initiative Contract and Its Subcontracts." National Cancer Institute contract provisions, office of contracts, NCI.)

NCI is not pursuing patents, and so this amounts to a nonpatenting policy. It runs a risk if products based on the genes discovered under CGAP require substantial subsequent development. Such inventions are likely because CGAP focuses on full-length genes of medical interest. The Bayh-Dole exemption was pursued because NCI could not rely on prevailing norms to ensure rapid availability of sequence data. If NCI could instead have relied on the proposed policy--public disclosure of DNA sequence soon after patent applications were filed--the Bayh-Dole exemption might have been unnecessary. Contractors could seek patents selectively or could file provisional patent applications, giving them a year to sort through the data in search of those gene sequences likely to require significant subsequent investment. This would achieve the end of early disclosure without the danger of eliminating all patents, even when they might be needed to induce investment in subsequent development.

The Single Nucleotide Polymorphism (SNP) Consortium has an unusual and sophisticated approach to keeping data in the public domain. The Consortium is funded by a group of 11 private firms collaborating with nonprofit research centers and the Wellcome Trust, to create a collection of sequence differences in the human genome. The SNP Consortium is pursuing a policy of open disclosure linked to patent applications. The Consortium applies for patents but only to establish priority dates and to secure standing as inventors (to contest other patents if necessary), but not to secure commercial patent rights. SNP Consortium sequence data are published, and patent applications abandoned, after a period sufficient to prevent others from filing patent applications using Consortium data, thus preventing encumbrances on open access.

The Rice Genome Project is another public-private consortium generating sequence data with great scientific value and commercial interest. The Internatoinal Rice Genome Sequencing Project partners have agreed to a policy of rapidly disclosing sequence data (but not necessarily data about annotation--gene location and function), similar to the Bermuda Accord. Monsanto provided draft genomic sequence data in April 2000. Monsanto does not charge for access to the data, but it requires users to sign an agreement that gives Monsanto rights to a nonexclusive right to any patents, thus ensuring nonexclusive licensing of any products resulting from use of its data. The Monsanto rice sequence access policies set precedents for Celera's policies for human genome sequence data.

Some privately funded research produces public domain data. Merck funded a Washington University project to sequence expressed sequence tags of genes and put them immediately into the public domain. It then created the private, nonprofit Merck Genome Research Institute (MGRI) to foster public domain science, to accelerate progress in genomics, and to ensure open access to critical research tools. MGRI permits its grantees to seek patents, but bars exclusive licensing. (The Institute is not currently accepting proposals, and its future is uncertain.)