Thoughts on Ex parte Khvorova

On July 13, 2015 the Patent Trial and Appeal Board (“PTAB”) issued a decision in Ex parte Khvorova (Appeal no. 2012-010359, appl. ser. no. 10/940,892). The PTAB affirmed the examiner’s rejection of the only independent claim at issue under 35 U.S.C. § 101 for being directed to statutorily ineligible subject matter. This is significant because it is, to my knowledge, the first PTAB decision on § 101 eligibility in the life sciences. Until now, the PTAB has been remarkably silent on eligibility in life sciences in the midst of significant judicial activity.

Claims

The invention relates to synthetically designed and produced small interfering RNA (“siRNA”) molecules, which were discovered in nature for their ability to suppress gene expression. Claim 85 recites a method of rationally designing siRNA molecules with optimal suppression efficacy as follows:

A method for obtaining an siRNA sequence for a target gene, wherein said siRNA sequence comprises 19 – 30 nucleotide bases, said method comprising the steps:

(a) selecting a target gene;

(b) generating a set of candidate siRNA sequences that are between 19 and 30 bases in length, wherein each of said candidate siRNA sequences comprises a sense region of 19 – 30 bases, wherein said sense region comprises a sense sequence of 19 bases that is at least 79% similar to a region of the target gene;

(c) applying an algorithm to said set of candidate siRNA sequences, wherein said algorithm comprises a set of at least four variables, wherein each variable has a value that is determined by a binary condition that is defined by the presence or absence of a base at a position within the sense sequence, and within said algorithm each variable is multiplied by a coefficient;

(d) determining a predicted relative functionality of at least two siRNA sequences within said set of candidate siRNA sequences, wherein said predicted relative functionality is measured by applying said algorithm to compute a value based on an application of said algorithm;

(e) selecting from said set of at least two candidate siRNA sequences, an siRNA sequence for gene silencing based on said predicted relative functionality, wherein a higher value generated by said formula is indicative of greater predicted functionality and said siRNA for gene silencing that is selected has the highest value from within said set of at least two siRNA sequences; and

(f) generating an output comprising said siRNA sequence for gene silencing that is selected, wherein said output is in a form that is readable by a user, thereby obtaining said siRNA sequence for gene silencing.

Analyzing the PTAB Decision

The PTAB’s decision is probably correct in its ultimate outcome, but contains cursory analysis and several troubling statements and mischaracterizations of the law.

Generalizing the Claimed Invention

One problem with the PTAB’s analysis, which is becoming nearly universal in § 101 decisions these days, is its over-generalization of the claimed invention. The decision begins by characterizing/generalizing the claimed invention as follows:

On this record, Appellants’ claimed invention involves (1) generating data related to candidate sequences based on a target gene for siRNA; (2) applying an algorithm to the data; and (3) determining, based on the result, whether a particular siRNA will result in gene silencing (see, e.g., Claim 85).

PTAB Decision at 6.

While Claim 85 is indeed very general, and the PTAB was likely correct in rejecting the claim on that basis as discussed below, it is not helpful for the PTAB (or the CAFC or patent examiners) to provide these opening general descriptions of what the invention “involves.” The PTO’s revised subject matter eligibility guidelines expressly retracted an earlier, widely criticized framework of scrutinizing claims merely because they “involve” a judicial exception. But the PTAB has done exactly that in this decision. As discussed below, this confirms concerns some expressed with the revised PTO guidance that the change from “involving” to “directed to” might not have much practical effect.

Even worse, creating an artificial generalized version of the claim at the outset is likely to doom the analysis. The claim does not merely recite “an algorithm,” but instead an algorithm with some level of detail. As the Ariosa and UURF decisions make clear, opening by characterizing the claim as “involving” “an algorithm” skews later analysis of the actual claim by reinforcing the tendency toward ignoring important details in the language of the claim itself.

Applying Ariosa

The Ariosa decision is very problematic on several levels. The PTAB decision here picks up on one of the worst of these and illustrates the mischief some predicted in the wake of Ariosa.

For example, the PTAB takes Ariosa’s superficial analysis of the first step in its Mayo analysis and goes even lower:

Consistent with Ariosa, because the ability of siRNA sequences to silence genes is “directed to naturally occurring phenomena, we turn to the second step of Mayo’s framework. ….”

PTAB decision at 7. The PTAB’s “analysis” of step one of Mayo is thus limited to one sentence. And even this one sentence illustrates a disappointing lack of depth. “[T]he ability of siRNA sequences to silence genes” is indeed a “naturally occurring phenomenon.” But claim 85 does not recite or purport to cover this phenomenon per se. Instead the claim recites a method for the rational design of artificial siRNAs with better ability to silence genes in human-made applications (e.g., therapeutics or laboratory research). That the PTAB summarily deems this “directed to” (as opposed to merely “involving”) the stated natural phenomenon with no analysis or explanation is worrisome.

This reinforces a concern some expressed after the PTO issued its revised subject matter eligibility guidance late last year. The Coalition for 21st Century Medicine, for example, noted gratefully that the PTO changed the threshold question from whether the claim “involves” a judicial exception to whether the claim is “directed to” an exception.  (See page 3). But the Coalition simultaneously voiced concern that this may end up being a distinction without a difference in practice. Id. (paragraph bridging pages 3 and 4). Here the PTAB used the words “directed to” but its analysis (or lack thereof) clearly shows it determined the claim failed step one of Mayo merely because it “involved” (i.e., was related to) a natural phenomenon.

Buzzwords and Superficial Analysis

Again taking a cue from the Federal Circuit in Ariosa and UURF, the PTAB’s analysis in this case is disappointingly shallow. One prime example of this is where the PTAB engages in the critical Mayo task of comparing the claimed invention to what was “routine and conventional” in the art. The PTAB’s consideration of this central question is limited to the following:

At the time of Appellants’ claimed invention, the analysis of gene sequences to allow researchers to predict whether a siRNA would silence genes was known in the art (see FF 1, 2, 4). The difference is that while the prior art recognized that sequence data can be analyzed to identify siRNAs, the prior art methods do not utilize the same algorithm as that disclosed in the Specification to manipulate the sequence data (FF 4). Cf. Parker v. Flook, 437 U.S. 584, 585 (1978) (“[t]he only novel feature of the method is a mathematical formula”).

Thus, the notion that sequence data relating to siRNA can be manipulated to allow an artisan to predict whether a particular siRNA will function to silence genes was “‘well known,’ to the point where, putting the formula to the side, there was no ‘inventive concept’ in the claimed application of the formula.” Mayo, 132 S.Ct. at 1299, citing Parker v. Flook, 95 U.S. 594.

PTAB Decision at 7.

The PTAB gives no specific differences between the prior art efforts at identifying functional siRNAs and the algorithm in claim 85. The full extent of the analysis is: “the prior art methods do not utilize the same algorithm.” There is no detailed comparison of the features of each process (prior art and claimed), but at this critical step in the eligibility analysis it would have been helpful to know whether the claimed process incorporates new data inputs, selects important minimal inputs from a list of candidates known in the art, gives different weight to known inputs or combines them in a new way, etc. We are instead left to take the PTAB’s word that (1) there is no difference between the rest of the claimed process and what was in the art and (2) there is some meaningful difference between the algorithm recited in the claim and the routine criteria for siRNA selection known in the art (which is difficult to imagine given how generic the algorithm in claim 85 really is).

Correct application of Mayo requires a detailed comparison of the claimed method to the prior art showing, step-by-step as well as whole method against whole method, how each compares to the other. The PTAB does not analyze the individual steps or how they integrate into the overall process, but instead boils everything down to the algorithm. In support of this approach it cites the portion of Flook that was until very recently widely considered to have been overruled by Diehr: isolate the algorithm in the claim and disregard it in comparing the method to the prior art. Parker v. Flook, 437 U.S. 584, 585 (1978) (“[t]he only novel feature of the method is a mathematical formula”).

Perhaps the worst example of cursory analysis is when the PTAB responds to one of the Appellants’ argument with a non sequitur. The Appellants argued that “all processes use general concepts, and use of a general concept does not render a claim ineligible for patent protection if the claim is directed to more than the abstract idea or general concept.” In other words, modification or adaptation of a general concept in some specific way to apply it to a specific context or application is eligible for patenting.

The PTAB avoids addressing this crucial point in the ongoing saga of patent eligibility (see Ariosa, UURF, etc.) by reciting unresponsive case law buzzwords without any discussion or analysis: “The ‘prohibition against patenting abstract ideas ‘cannot be circumvented by attempting to limit the use of the formula to a particular technological environment.’” This quote misses the point because, unlike the patentee in Flook, Appellants never argued that taking a general abstract idea, which could conceivably be applied as-is in numerous fields of technology, and limiting the claim to the use of that still general abstract idea in a specific field is sufficient for eligibility. Instead, Appellants made the far more compelling argument that taking a general idea and devising a specific modification of it such that it performs better or is better suited to a new application is sufficient for eligibility. Knowing the PTAB’s considered view on this important battleground of eligibility would have been helpful. Instead we are left to guess whether the PTAB simply missed the point or in a confusingly imprecise way rejected it.

Imprecise Discussion/Application of Governing Law

Because this is the first PTAB decision on § 101 eligibility in the life sciences, it has the potential to significantly impact examiner behavior. As such, I hoped for more precision in the PTAB’s language to avoid overstatements of the law or other excerpts that could be misinterpreted by the examining corps.

For example, at one point the PTAB states: “Unlike the claims at issue in Diehr, Appellants’ claims do not involve the transformation of an article into a different state or thing. See Diamond v. Diehr, 450 U.S. 175, 184 (1981).” It is true that Claim 85 does not appear to recite transformation of a physical article. But the phrasing of this statement casts Diehr in a broad negative light and may be viewed as resurrecting machine-or-transformation, rejected by the Supreme Court as the exclusive test for eligibility in Bilski.

This statement could be read by examiners, struggling to find any clear landmark through the sandstorm of § 101 eligibility, to suggest such a transformation is required for eligibility. If instead the PTAB’s intention was simply to point out a factual distinction between Diehr and the present case—i.e., Appellants can’t rely on Diehr due to important factual distinctions—this could have been better stated as follows:

In assessing eligibility, it is helpful to compare the present claims to those courts have found eligible and those found ineligible. While Bilski teaches a transformation is not required, recourse to Diehr is not necessarily helpful to Appellants’ claims because, unlike the claims at issue in Diehr, Appellants’ claims do not involve the transformation of an article into a different state or thing. See Diamond v. Diehr, 450 U.S. 175, 184 (1981).

Getting It Right, But Possibly Too Late

Ultimately the above failings in the PTAB decision are most disappointing because they were unnecessary. At the very end of the decision, the PTAB finally dives into the details of and actually analyzes the claim language. The PTAB finally gives a cogent reason for rejecting claim 85 that could have made this decision about two paragraphs long and avoided all the potential collateral damage noted above.

Claim 85 does not recite a specific algorithm, and preempts any algorithm for designing siRNA sequences. The claim limitation recited by Appellants regarding the four asserted “variables” can simply represents [sic] the presence or absence of the four alternative bases possible in siRNA, adenine, guanine, cytosine, and uracil (FF 5). The claim requires no particular coefficient for the algorithm, nor does the claim impose any specific metric to identify improved gene silencing. Ariosa notes that “patent claims should not prevent the use of the basic building blocks of technology—abstract ideas, naturally occurring phenomena, and natural laws.” Ariosa at 7. Instant claim 85, by failing to recite any specific elements of the claimed algorithm, specifically attempts to fully preempt any process for designing siRNA.

PTAB Decision, paragraph bridging pages 8 and 9 (emphasis added). The focus on preemption is worrisome, especially in view of Ariosa’s arguable rejection of preemption as the test for eligibility.

But at least the PTAB finally confronts the fatal problem with this claim. It is directed to an abstract idea—i.e., using an undefined number of unspecified variables, together with one variable simply corresponding to which of the four naturally-occurring nucleotides is present a particular location, to calculate a score that in some undefined way predicts the ability of any given siRNA to silence any target gene. The specification discloses numerous more specific examples of such an algorithm, but in this application the Appellants chose to pursue this broadest possible conception of the method. This does indeed appear to be a claim improperly directed to an abstract idea under Alice and Mayo. But none of the superficial, misdirected and misapplied “analysis” in the preceding seven pages of the PTAB decision was necessary. This final point is sufficient.

I don’t want to make a mountain out of a molehill. After all, the PTAB appears to have reached the right conclusion here. If this were just any PTAB decision I would likely just be grateful for that. But this is the first PTAB decision in molecular biology since patenting in that field got turned upside down. Practitioners and examiners alike are hungry for guidance. That’s why this decision has significant potential to set the tone for more PTAB decisions and for examiner rejections. This case shows the importance of the PTAB drafting opinions that succinctly state the clearest, independently sufficient basis for rejection and avoiding unnecessary additional commentary that may lead examiners to unfounded future rejections of important biotech patent claims.