Of the many issues broached in relation to any patent or application, few cause as much hair pulling and hand wringing as the written description requirement. For years this requirement that "the specification shall contain a written description of the invention" was analytically lumped with another requirement, the enablement requirement, which stipulates that the application must describe how to make and use the invention so that one "skilled in the art to which it pertains" may do so. In fact, for many inventions it is difficult, if not impossible, to fulfill the enablement requirement without also meeting the written description requirement. Neither the courts nor the Patent and Trademark Office frequently felt obligated to parse the details of these requirements. When the patentability of altered living organisms was affirmed in Diamond v. Chakrabarty , the wave of biological applications and patents that followed was like a tsunami, uprooting long-established ideas about written description and requiring a great deal of rebuilding and restructuring in its wake.
The foundation for our present understanding of the written description requirement for genomics patents was laid in 1991 by the Federal Circuit in Fiers v. Sugano , when the court indicated that the specification must convey to one skilled in the art that the inventor had "possession" of the invention as claimed at the time the application was filed. Applying this broad requirement to claims for a DNA molecule, the court went on to say that "possession" must be demonstrated by a description of the DNA itself, not a method for isolating it.
This reasoning explains how one may enable an invention, but still fail to provide a written description. An application can describe how to make and use a DNA molecule by providing detailed instructions for its isolation, yet it may provide little to no description of the molecule itself. The "possession" language judicially incorporated into the written description requirement merely sets a standard for judging what characteristics are sufficient to describe the invention.
This "possession" standard was explored further in relation to DNA molecules in University of California v. Eli Lilly & Co. The specification in that case only described a cDNA for rat insulin, yet other mammalian and vertebrate cDNAs for insulin were claimed, including human. Though the specification provided directions for obtaining these other cDNA molecules, the court found that it did not describe enough of their characteristics to show "possession" of them, so the claims to all mammalian and vertebrate insulin cDNAs were held invalid.
The types and numbers of characteristics required to show "possession" have recently been addressed within the Patent and Trademark Office in a set of Examiner Guidelines on Written Description http://www.uspto.gov/web/offices/com/sol/notices/writdesguide.pdf These Guidelines organize the analytical structures that have been rebuilt in the aftermath of the biological inventions wave. They do not have the force of law, but they do represent Patent and Trademark Office policy and should be complied with by the Patent Examiners. It is important to keep in mind, however, that courts may decide that the Guidelines themselves are too broad, or that they are too liberally applied, and may still find a patent invalid for lack of written description despite its compliance with the Guidelines.
The Three Ways to Show Possession
The Written Description Guidelines indicate that there are three basic ways to show possession of an invention. These methods are not specific for genomics inventions, so some are more applicable in that field than others. First, the specification may show actual reduction to practice. In the case of a claim to a DNA molecule, this would entail isolating and sequencing the molecule and providing other available identification. In fact, because there are so few identifying characteristics of a DNA molecule other than its sequence, this may prove to be the only way to show possession of a DNA molecule that is reliable in court.
The second method is through a clear depiction of the invention through detailed drawings. This method may not be useful for most genomics or proteomics inventions because the information needed to render a drawing will already satisfy the requirement under the first or third methods. However, it may have some applicability in proteomics inventions where most information is structural.
The third method of demonstrating possession involves providing a description of sufficient relevant identifying characteristics of the invention. The Guidelines provide some examples of such characteristics, and more information regarding what is sufficient and what is not may later be included in Examiner Training Materials. In one example, the Guidelines indicate that a restriction map of an isolated DNA molecule may be sufficient to show possession, but evidence that it can be digested with a nuclease is not because almost all DNA can be so digested. Including a partial sequence may also be helpful. Similarly, evidence of a protein that is produced by a claimed mRNA molecule is strong evidence of possession of the mRNA molecule.
Given the ease with which an isolated nucleic acid can be sequenced, the third method for demonstrating possession is more useful for protein claims. In a more traditional scenario, where a protein is isolated, but not sequenced, the protein may still be claimed if possession is demonstrated by a combination of other characteristics such as a partial sequence, structure, binding affinity, binding specificity, molecular weight, length, unique cleavage by enzymes, isolectric point of the molecule or its fragments, comparision of enzymatic activity and antibody cross-reactivity.
A more recent issue is whether a protein can be claimed based upon possession of a DNA molecule. Such a claim may be possible without ever isolating the protein if enough of the above characteristics can be predicted from the raw sequence through homology and other known factors. However, there is no guidance from the Patent and Trademark Office or in the caselaw as to how many characteristics should be described in such a situation and how certain the predictions must be. There is also the possibility that, even if sufficient predictions are made to demonstrate possession, and a patent issues, the protein as claimed with the limitations based on homology will not, in fact, resemble the protein encoded or produced by the DNA. Thus, the patentee may be left with a commercially valueless patent.
As the examples above make clear, the Written Description Guidelines are useful in providing some organization of the varied caselaw regarding written description. They give inventors and companies a way of judging when a patent application should be filed and what it should include and may also reflect on whether an issued patent is valid.
Rochelle K. Seide is a partner at the law firm of Baker Botts, where she specializes in biotechnology, intellectual property, and patent issues. She also has a PhD in human genetics. Michelle LeCointe is an associate at Baker Botts and a registered patent agent. They can be reached at [email protected] and [email protected] .
TrendSpotter is a weekly column that focuses on how trends in politics, patent law, and the US and European markets will affect the genomics industry. The column appears every Friday. Next week Ira Leiderman, managing director of the Palladin Group, will write about ways to raise money in a market downturn.
To access previous columns just enter the word "Trendspotter" in the archive search window on the homepage.