Patent Office Catch-22

    By Kevin E. Noonan --

There was a time not so long ago where the U.S. Patent and Trademark Office seemed to be sincere in trying to accommodate the needs of inventors and the public in providing fair examination of cutting-edge technologies.  This is particularly true in the biotechnology arts, with such innovations as the sequence listing database and (after some resistance) recognizing that nucleic acid claims were not prima facie obvious.  That has changed, with the Office seeming to be happy in many instances to play an advanced game of "gotcha" with its biotechnology customers.

This new attitude was illustrated once again in a presentation by Julie Burke, a "Quality Assurance Specialist" in Group 1600 at the Biotechnology/Chemistry/Pharmaceutical Customer Partnership meeting held at the Patent Office this Wednesday.  In her presentation she addressed "incorporation by reference," the sufficiency of disclosure requirements, and the standard for determining whether clarification of the identity of certain nucleic acids would be new matter.  The particular issue involved whether and under what circumstances it was sufficient to identify a nucleic acid by its GenBank Accession Number.  Ms. Burke sets out the "problem" in the very first substantive slide:

In patent specifications, every element or ingredient of the product should be set forth in positive, exact, intelligible language, so that there will be no uncertainty as to what is meant.

The relationship between sequence information submitted to a public database and the amino acid or nucleic acid product it identifies is sometimes indefinite, uncertain, and arbitrary.

She goes on to identify the bases for the "uncertainty":  sequence information in public databases can "change from time to time," and these changes can be made by third parties.  These changes can constitute new matter, according to Ms. Burke, and implicate the Office's "New Matter Guidance" to Examiners to sensitize them to the issue.  These considerations implicate the requirements for incorporating material by reference, which also is subject to new matter prohibitions, i.e., sequences that have been changed after an application incorporating them by reference has been filed.  The rationale is that only a sequence having one version has been "uniquely identified"; even prior to the filing date, when multiple sequences for the same gene are identified, the existence of these sequences means the gene "may not be uniquely identified."

Many of the examples in Ms. Burke's presentation involve claims to isolated nucleic acids, and insofar as the application of the "guidelines" are limited to such claims they may have some justification.  But the reality is that the 25,000-30,000 human genes found by the Human Genome Project have been disclosed in public databases for almost 10 years, and patent applications on these genes were filed at the turn of the century.  Ms. Burke's presentation also mentions that the Office has changed its former policy of examining up to 10 sequences in a single application, and will now only examine one sequence per application (OG Notice dated March 27, 2007; see Patent Docs report).  As a consequence, only a small percentage of human genes have been patented to date, and the likelihood is that the 20-year patent term limits will preclude most of the genes disclosed in most of the applications from ever being the subject of granted claims.  Moreover, the business and intellectual property of the next 50 years in biology will not be the identification of genes but rather how different genes are differentially expressed in disease (inter alia, for diagnostics) or in response to different physiological stimuli such as therapeutic drug administration.  These experiments take advantage of the information about these genes, using array technologies and other methods.  The issue thus is not the same as with claims to the genes themselves; all that is required is that the skilled worker understand which genes are differentially expressed.  Sequence variations, whether the results of allelic variation or sequencing disparities from different research groups thus would not create any ambiguity for the skilled worker as to the identity of the genes.  However, the current Patent Office focus on whether there are sequence variations that preclude these genes from being "uniquely identified" elevates the formalism of sequence identity (important for sequence claims) to another "bright line" rule having the potential for being indiscriminately applied.  And the proposed solution, including the sequence of genes in a sequence listing, merely increases the applicant's burden without improving the quality of the disclosure for known genes.

Which, unfortunately, is not a unique position for the Office to take in recent years.  Perhaps a change of administration will return the Office's attitude from confrontation back to cooperation.

Comments

[Michael]

I agree. It is unpractical and prohibitively expensive to require applicants to provide sequences for all the genes, for example in a gene expression profile. With unigene ID's in GenBank, the revisions, if any, may not alter the claim scope. I don't understand Ms. Burke's rationale.

[Chris Singer]

Kevin and Michael:

In my opinion (and as noted in Kevin's post) the PTO's motivation for raising this issue derives primarily from applications with claims directed to a sequence that is only identified by GenBank number in an application. By the time prosecution is underway, the GenBank sequence information may have been modified from its original version, even though the GenBank number remains the same. While this can confuse things, GenBank provides easy-to-find links to prior versions of sequences relating to GenBank Accession numbers that have had sequence information updated over time. Thus, a reasonable Examiner and/or one of skill in the art can determine (and demonstrate) what the sequence information was at the time of filing.

While this creates a bit more burden on applicants and practitioners, applicants having claims drawn to nucleic acid sequences (with only reference to GenBank Accession numbers) should be made aware of this potential hitch and be prepared to deal with it.

Whether the PTO will attempt to distinguish applications having claims that are sequence-specific from those that are non-specific is another question...

Chris

[Kevin E. Noonan]

Dear Michael and Chris:

Insofar as the "rule" is applied solely to claims like "An isolated nucleic acid having a sequence identified by GenBank Accession No. XXXXX," the Office has at least a modicum of logical basis for the rule; for example, if the claim is asserted against an alleged infringer, which sequence do you choose (if there are multiple ones).

The post is concerned with the tendency of the Office to make these "guidelines" into bright line rules, and I can envision an Examiner saying that a claim to an assay for detecting gene expression modulation of a handful of genes, which are identified by GenBank Accession No. in the specification, are not "uniquely identified." Which is ridiculous but not unimaginable.

Thanks for the comments.

[nadiamac]

Do you have copies of the handouts from the presentation, or know where these might be found?

[babaloo]

Considering that many of the genes in GenBank came from one person (specifically from Dr. Craig Venter), the chance of variations and particularly of more representative forms of genes occurring in the future is highly likely. This could cause severe problems with incorporating genes from GenBank by reference, even though that may not be a patentable part of the invention.
Also consider: Her requirement of such a high level of description and specificity is not appropriate in any part of the patent office, as she seems to be requiring the complete specifications of everything, which long has been recognized as unnecessary.

[babaloo]

Oh, for the good old days, when the biotech unit had ombudsmen to facilitate appropriate patenting, instead of QA specialists who seem intent on slowing down innovation in America!

[Chris Singer]

Dear nadiamac-

You can find copies of the presentations at http://www.cabic.com/bcp

[Jeff]

This is a problem for older cases, but now GenBank accession numbers have a version number. This means no indefiniteness or ambiguity if the sequence later changes. I asked Ms. Burke (at one of the PTO road shows) about disclosing an accession number with a version identifier, but she still felt that this would not be a unique identifier - sounds like a bright line rule already, as Kevin suggests.

[Kevin E. Noonan]

Dear Jeff:

In this case, I wish you could have told me I was wrong.

Thanks for the comment (and the info).