By Donald Zuhn --
As we reported yesterday, the Federal Circuit heard oral argument today in In re Kubin (see "In re Kubin to Be Argued before the Federal Circuit on Thursday"). Kevin Noonan will be providing a complete analysis of the oral argument in this case in a subsequent post. One interesting aside, however, involved an exchange between the Court (specifically Circuit Judge Rader) and USPTO Associate Solicitor Janet Gongola regarding Example 11 of the new Written Description Training Materials.
Last March, the Patent Office announced that it had updated the training materials to be used by examiners in the examination of patent applications for compliance with the written description requirement. The revised training materials replaced the previous set of training materials issued by the Office in 1999. Perhaps the most interesting of the seventeen examples in the new training materials is Example 11, which concerns claims that are directed to a polynucleotide or polypeptide sequence that shares percent identity with another sequence (see "An Analysis of the New Written Description Training Materials - DNA Hybridization & Percent Identity").
Example 11 is divided into two sections, one in which there is no art-recognized structure-function correlation for the claimed sequence (Example 11A), and one in which there is an art-recognized structure-function correlation for the claimed sequence (Example 11B). Both subparts present identical exemplary claims reciting an isolated nucleic acid that encodes a polypeptide with at least 85% amino acid sequence identity to SEQ ID NO: 2, and nearly identical exemplary claims reciting an isolated nucleic acid that encodes a polypeptide with at least 85% amino acid sequence identity to SEQ ID NO: 2; wherein the polypeptide has a recited activity. The only difference between the hypothetical specification of Example 11A and the hypothetical specification of Example 11B is that the latter identifies two domains -- a binding domain and a catalytic domain -- that are critical to the recited activity. The training materials specify that the first exemplary claim in each subpart satisfies the written description requirement. The training materials also state that while the second exemplary claim in Example 11B satisfies the written description requirement, the second exemplary claim in Example 11A does not.
Last June, during a presentation at BIO 2008, Group 1600 Director Dr. George Elliott acknowledged that Example 11 represented a reversal in the Office's position (see "Docs at BIO: Representatives from JPO, EPO, SIPO, and USPTO Discuss Recent Developments in Japan, Europe, China, and the U.S."). In particular, Dr. Elliott noted that the Office's position had previously been that claims lacking functional language (such as the first exemplary claim in Examples 11A and 11B) failed to comply with the written description requirement and that claims possessing such language (such as the second exemplary claim in Examples 11A and 11B) complied with the requirement. Dr. Elliott contended that claims with functional language (where the specification lacks any teaching regarding the amino acid residues that are tolerable to change) do not satisfy the written description requirement because "the minute you add function, you've limited the claim to a subset of species, and you don't know which species are in the subset and which species aren't." In other words, absent any teaching of structure/function relationships, the Patent Office's position is that one cannot determine the species that share at least 85% sequence identity with the recited sequence and also possess the recited function. For a more thorough discussion of the rationale behind the USPTO's reversal on functional limitations, Dr. Elliott recommended that attendees read Ex parte Porro, which involved an application filed in June of 2003 (for a discussion of the Porro decision, see "The Written Description Training Materials and Ex parte Porro").
With respect to today's oral argument in Kubin, the Court speculated that the Patent Office may have reversed its position on the fact pattern set forth in Example 11 in order to support its argument that Kubin lacked an adequate written description despite the recitation in claim 73 of an activity limitation (i.e., binding CD48). A transcript of the exchange between the Court and Associate Solicitor Gongola follows:
Gongola: I'd like to move now -- if the Court has no more questions on obviousness -- into the written description area. This Court found that written description was not satisfied under the only two tests that Kubin ever raised before the Board: the representative number of species test and the structure-function correlation test. Now, the Board found, at Finding of Fact 20 and 21, that the specification only taught how to make cDNA molecules that encode NAIL identically. That's one subgenus. The Board found, at Finding of Fact 22, that the specification did not teach anything about any variants.
Court: What about the conservative substitutions set forth at page 63?
Gongola: The language on conservative substitution, using the Board's own findings, does not demonstrate -- Finding of Fact 23 -- which 20% of the amino acid residues should be changed to maintain function. So that discussion of conservative substitution may teach how to make the variants, but it doesn't teach what those variants are. It doesn't describe them so that a person of skill in the art would understand that Kubin possessed the variants. There -- the Board Finding of Fact 25 says that there are a very large number of modifications that can be made. One in five amino acids can vary. We have no idea which of the 21 to 220 -- 22 to 221 portion of NAIL to change and maintain binding -- which amino acids have to remain and which can be substituted. But more than that, even if this Court would accept the discussion that the conservative substitution could somehow provide written description support for variants, it wouldn't still do so for the full scope of the genus. That discussion would only apply to variants made by substitution. Applicant -- Kubin has defined a variant to be made by a substitution, an insertion, or a deletion. Conservative substitution dialog doesn't speak to anything about variants made by insertions, or variants made by deletions. So therefore, Applicant Alonso still has -- I'm sorry, Kubin . . .
Court: Yes, but the description of the variations is also linked to a protein having activity Y. Therefore, this satisfies the function-structure alternative test for written description, doesn't it?
Gongola: No, your Honor. The function-structure test requires an identification of a structure common to all members of the genus. We do not have that identification of a common structure. Twenty percent . . .
Court: I'm quoting almost out of PTO's manual on written description -- verbatim -- when I give my example there, am I not? The variations plus the activity Y -- protein having activity Y -- isn't that exactly the PTO's manual on written description?
Gongola: Are you referring to the training materials?
Court: Yes, I am.
Gongola: That language may be found in the training materials, but it is not -- it's guidance. It's -- each case of written description has to be decided on its own facts.
Court: The Patent Office is guiding applicants on how to do things wrong?
Gongola: No, your Honor, but guidance provided in a training document cannot be taken and applied to each case. Each case has to be decided on its own facts.
Court: So it's just guidance to examiners on how to do it wrong?
Gongola: No, your Honor -- respectfully -- it is guidance to the public as well, but . . .
Court: But if we take that guidance, we don't get to your result, do we?
Gongola: No, your Honor, I respectfully disagree; we do. In Carnegie Mellon, this Court has explained that when there's substantial variation within a genus, an applicant has to describe a sufficient number of species to reflect the variation. Kubin has not done that here. Kubin hasn't described any variation. If we want to look at the training materials, the Board only . . .
Court: It's interesting that this was revised immediately after Kubin -- the Kubin result -- wasn't it?
Gongola: That . . .
Court: Which is kind of an admission that the example did track Kubin and was detrimental to your position, wasn't it?
Gongola: No, your Honor.
Court: So, despite your smiling defense, the facts tend to give us a different conclusion.
Gongola: No, your Honor, that's not correct. The training materials were not revised post-Kubin to somehow capture Kubin. The revisions have been in the works for a very long time.
Court: I see, okay.
Gongola: So it's just coincidence that Example 11 may seem to look like the Kubin fact pattern. But if we actually look at Example 11, it is distinguishable from the facts here. Example 11 goes to a nucleic acid that encodes a polypeptide that has certain homology -- 85% homology -- and a certain activity. Now here's where -- that sounds a lot like Kubin's claim. I agree with that. But here's where the difference resides: the specification in Example 11 disclosed two specific domains that were responsible for the activity. Kubin's specification here doesn't contain any similar disclosure. So the fact pattern here is different from the fact pattern in the revised training materials. But I also want you to know -- these materials, as you point out, were not even in existence when the Board rendered its decision. So it really isn't proper to be considering them right now, since the Board didn't have a chance to consider them in the first instance. And on top of that, they're only guidance. They're not a rigid rule.