Monday, August 15, 2011

AMP v. PTO Does Not Appear to Invalidate Very Many Gene Patent Method Claims

In this post, I explain why my recent analysis of 533 patents identified by Jensen and Murray as "gene patents" (described in a previous post) suggests the Federal Circuit’s recent decision in AMP v. PTO implicates the validity of very few gene patent method claims.

Background

As I pointed out in a blog post last December, the patent eligibility of Myriad's claims to methods of detecting mutations in the BRCA gene by "analyzing" or "comparing" DNA sequences hinged entirely on the court's interpretation of the word "sequence" as it appears in the claims. In the field of molecular biology, the term "sequence" is routinely used not only to refer to the description of the chemical structure of a DNA molecule, but also to the actual DNA molecule as well.


For example, a scientist might say that "I determined the sequence of the BRCA gene." In this context, she is using the word sequence in reference to the abstract description of the order of nucleotides appearing in the gene. But the same scientist might also say that "I cloned the BRCA sequence into an expression vector," in which case she is using the term "sequence" to refer to the DNA molecule itself.

In the first example, "sequence" refers to abstract information describing a chemical structure, while in the second "sequence" is used to refer to a physical object, i.e., a specific DNA molecule. Molecular biologists are usually not bothered by the dual meaning of "sequence," because in practice the meaning of the term is understood based on the context in which it is used. However, it can lead to ambiguity, as seen in AMP v. PTO.

In AMP v. PTO, Myriad argued that as used in its claims the term "sequence" refers to an actual DNA molecule, and thus that analyzing or comparing "sequences of the BRCA gene" would involve physically manipulating and processing molecules in a manner that would render the claims patent eligible under Prometheus. I think it is clear that if the court had adopted this interpretation of "sequence" it would have upheld the validity of the claims. The ACLU plaintiffs never argued that a method claim that requires the analysis of physical DNA molecules would be patent ineligible.

Instead, as explained in a previous post, the ACLU's case was based on its argument that the term "sequence" in the claims refers to information, not to a physical molecule, and hence would cover the mental activity of analyzing or comparing genetic information. The court adopted the ACLU's interpretation of "sequence,” and ruled the claims patent ineligible for claiming nothing more than a mental step. This outcome was specifically dictated by the Federal Circuit’s earlier Prometheus decision, where it stated that a claim directed only to mental analysis of information is patent ineligible.

The Federal Circuit based its interpretation of “sequence” on its finding that Myriad's patent specifications implicitly defined “sequence” broadly to cover pure information, as set forth in this excerpt from the case:

The patent specifications make clear that “sequence” does not exclusively specify a DNA molecule, but refers more broadly to the linear sequence of nucleotide bases of a DNA molecule. For example, Figure 10A–10H is described as showing the “genomic sequence of BRCA1.” ′473 patent col.5 l.66. Figure 10 does not show a physical DNA molecule; the figure lists a series of letters (Gs, As, Ts, and Cs) corresponding to the nucleotides guanine, adenine, thymine, and cytosine of a DNA molecule. Similarly, the patent specifications state that “[t]he nucleotide sequence for BRCA1 exon 4 is shown in SEQ ID NO: 11.” Id. col.53 ll.50–53. SEQ ID NO: 11 again lists a series of Gs, As, Ts, and Cs corresponding to the nucleotide sequence of BRCA1 exon 4.

Thus, Myriad’s claims might very well have been found patent eligible if the specification had defined the term "sequence" to refer only the DNA molecule itself. Alternatively, the claim probably would have been found patent eligible if it recited analyzing "DNA molecules" instead of analyzing sequences. If the claims could only be infringed by someone who physically analyzes the DNA molecule, it would certainly be patent eligible under the Federal Circuit's interpretation of Bilski as set forth in Prometheus.

Claims that could be infringed by merely analyzing genetic data appear to be rare in Jensen and Murray dataset

In essence, Myriad’s method claims were patent ineligible because, as interpreted by the court, they could be infringed by analysis of genetic information. However, my research suggests that few if any of the gene patents identified by Jensen and Murray fall into this category. To the contrary, the vast majority of the claims in these patents would appear to be patent eligible.

Of the 533 patents I analyzed in my study, I only found 12 that included a claim reciting a method of analyzing a DNA sequence for a mutation or variation, and most if not all of those claims appear to require a physical manipulation of a patient's DNA in order for there to be infringement.

In 8 of the 12 patents (6,395,482; 6,087,107; 6,458,541; 6,743,579; 5,830,649; 5,840,486; 5,955,265; and 6,410,226) the broadest claims specifically require obtaining a sample from a patient's body, or analyzing for the genetic variation directly in a patient's body.

For example, Claim 1 of US patent number 6,395,482 recites:

1. A method for determining susceptibility in a human subject to schizophrenia wherein the method comprises the steps of:

(a) removing a bodily sample from the subject, wherein the sample comprises a polynucleotide sequence of a PRODH gene;

(b) determining whether the PRODH gene of the bodily sample comprises a DNA sequence comprising a variation in SEQ ID NO:1 consisting of a T to C transition in the first position of codon 497, such that the presence of said variation in said PRODH gene is indicative of said subject's susceptibility to schizophrenia.
These eight claims would all appear to be patent eligible under AMP and Prometheus because they all involve physically manipulating a human sample obtained from patient.

The remaining four of the 12 patents (5,916,748; 6,630,304; 5,989,815; 6,432,644) are a little more ambiguous, and do not explicitly recite obtaining or analyzing a bodily sample. But arguably all of these claims do require physical manipulation of a sample, depending upon how a court interprets the claim, which will depend in part on how the claim terms were used in the patent specification

For example, Claim 1 of 6,432,644 recites:

1. A method for diagnosing the presence of a polymorphism in human KCNE1 (the coding region of which is bases 193-579 of SEQ ID NO:3) which causes long QT syndrome wherein said method is performed by means which identify the presence of said polymorphism, wherein said polymorphism is one which results in the presence of a KCNE1 polypeptide of SEQ ID NO:4 with an altered amino acid, said altered amino acid being selected from the group consisting of: a) a Leu at residue 74.
Note that this claim uses "means plus function" language, and would probably be interpreted to only cover means for identifying polymorphisms (a polymorphism is a genetic variation) that are described in the specification. If the only means for identifying polymorphisms described in the specification rely on physically analyzing DNA molecules, the claim would appear to be so-limited and thus patent eligible.

Similarly, Claim 1 of 6,630,304 recites:

1. A method of diagnosing a susceptibility to osteoporosis in an individual, comprising detecting a polymorphism in a human BMP2 gene of SEQ ID NO: 1, wherein the presence of a "T" at nucleotide position 11980 is indicative of a susceptibility to osteoporosis, compared with an individual having an "A" at nucleotide position 11980.
This claim refers to detecting a genetic variation in a "gene," as opposed to a "sequence" (the language used by Myriad). If a court were to interpret the claim limitation "detecting a polymorphism in the gene" as requiring actual analysis of a DNA molecule (which could depend upon how the term "gene" is using the specification), the claim would appear to be patent eligible. However, if the claim is interpreted to encompass detecting a polymorphism by analyzing genetic sequence data, it would appear to be patent ineligible.


Monday, August 8, 2011

Will Gene Patents Impede Whole Genome Sequencing?: Deconstructing the Myth That 20% of the Human Genome Is Patented


I am in the process of finalizing an article entitled “Will Gene Patents Impede Whole Genome Sequencing?: Deconstructing the Myth That 20% of the Human Genome Is Patented,” which I think many readers of this blog will find of interest, particularly in light of the recent decision in the Myriad a gene patent case (AMP v. PTO). I will be presenting my work this Thursday at the Intellectual Property Scholars Conference at Depaul Law school in Chicago, and have posted a working draft on SSRN.

In my article, I point out that there is a widely held perception that 20% of human genes are "patented," and that these patents preclude researchers and clinicians from using, studying or even "looking at" the patented genes without obtaining a license or risk being sued for patent infringement. In recent years the basis for this belief has increasingly become obscured.

For example, in AMP v. PTO Judge Lourie states that it “is estimated that . . . By 2005, [the PTO] had granted 40,000 DNA-related patents covering, in non-native form, twenty percent of the genes in the human genome.” In support of this assertion he cites to a law review article: J. Rogers, Can You Patent Genes? Yes and No, 93 J. Pat. & Trademark Off. Soc’y 19, 40 (2010). If you read the article by Rogers, you see that he cites to other secondary sources for the proposition that 20% of human genes are patented. Ultimately, if you trace the string of references back to their primary source, one arrives at the seminal article by Kyle Jensen & Fiona Murray: Intellectual Property Landscape of the Human Genome, 310 Science 239 (2005).

It is no surprise that Judge Lourie believes that 20% of human genes are “patented” - over the years the figure has been quoted so often in so many venues that it has become something of an urban legend. Oftentimes the statement that 20% of human genes are patented is made without any supporting reference (for example, this occurred in a recent article published in Science entitled: The Human Genome (Patent) Project). Other authors, such as Judge Lourie, cite to various secondary sources that parrot the conventional wisdom. But if you trace back through the citations, invariably you will find that the sole basis for the myth that 20% of human genes are patented is the 2005 Science article by Jensen and Murray.

Believing that 20% of human genes are patented, and that a patented gene is effectively off-limits, many have become concerned that whole genome sequencing will result in the infringement of hundreds or even thousands of patents. This specific point was made by Judge Bryson in his dissent to AMP v. PTO. During oral arguments, he asked Myriad's attorney whether Myriad's isolated DNA sequence claims would be infringed by personal whole genome sequencing, to which the attorney essentially responded "no" (the ACLU attorney later voiced his opinion that the patents would cover genome sequencing).

Later during oral arguments, Judge Bryson expressed his concern that whole genome sequencing could result in infringement of thousands of gene patents. His perception that isolated DNA claims create a thicket of patents that will impede personal whole genome sequencing might have contributed to his view that these claims should be declared patent ineligible.

In fact, a careful reading of the supporting online material for the Jensen and Murray article reveals that their study has been grossly misinterpreted by those who claim the 20% of human genes are patented. In fact, those authors only purported to show that, with respect to 20% of human genes known at the time they conducted their study, either (1) the DNA sequence of the gene, or (2) the amino acid sequence encoded by the gene, was mentioned in a US patent claim. The myth that 20% of human genes are “patented” has taken root because too many have incorrectly inferred that the mere “mention” of a gene in a patent claim precludes all uses of the gene.

To better understand the actual implications of Jensen and Murray’s findings, I analyzed the claims from a random sampling of 533 of the 4270 patents identified in their article as “gene patents.” (Jensen and Murray graciously provided me with the entire data set of patents identified in their study). Significantly, I found that, under any reasonable interpretation, 140 of the 533 patents would not be infringed by any form of genetic testing. In fact, many of these patents only claim proteins, or recombinant cells, or methods for using DNA that have nothing to do with DNA sequencing. The implicit assumption that the “gene patents” identified by Jensen and Murray cover all uses of the gene (or in some cases corresponding protein) mentioned in the patent claims is clearly false.

The remaining 393 patents include claims with respect to which I cannot entirely rule out the possibility that at least some form of genetic testing would be found infringing. These claims fall into two categories - products claims directed to polynucleotides (e.g., DNA molecules), and method claims that might cover at least some forms of genetic testing. The language used in these claims is extremely heterogeneous, and it is impossible to predict with any certainty exactly how broadly a court would interpret their scope if they were ever asserted in litigation, but to varying degrees a majority of these patents would appear not to be infringed by at least some, perhaps all, forms of genetic testing.

Perhaps most significantly, few (if any) of these patents would appear likely to be infringed by next-generation whole genome sequencing technologies, particularly those that do not require DNA amplification.

In short, there is absolutely no basis to infer from the Jensen & Murray article that 20% of human genes are off-limits to the patents, nor that whole genome sequencing, and other multiplex genetic diagnostic testing technologies, would result in the infringement of a large number of human gene patents. To the contrary, it appears that a vast majority of these patents were drafted in a manner that would not encompass whole genome sequencing.

Monday, August 1, 2011

AMP v. PTO Casts Doubt on Patent Eligibility of "Purified" (as Opposed to "Isolated") Biomolecules

On January 5, 2001, the US PTO published Utility Examination Guidelines explaining its long-standing policy of treating "isolated and purified" DNA molecules as patent eligible. Throughout the Guidelines, the PTO consistently refers to these patent-eligible DNA molecules as both isolated and purified, but never explicitly attributes distinct meaning to the two terms. Instead, I think most people have interpreted the terms (as used by the PTO in this context) as essentially redundant, similar to someone referring to a contract provision as "null and void."

If anything, the PTO guidelines might be interpreted as treating "purification" as a more demanding requirement than "isolation." For example, at one point the guidelines state that "an inventor's discovery of a gene can be the basis for a patent on the genetic composition isolated from its natural state and processed through purifying steps that separate the gene from other molecules naturally associated with it,” perhaps implying that mere isolation will not be sufficient unless the DNA molecule is not subsequently purified through processes that separate the gene from other molecules.

In practice, it is my experience that the PTO and most practitioners have interpreted "purification" and "isolation" as essentially redundant. While most "gene patents" recite DNA molecules that are "isolated," "isolated and purified," or "recombinant," there are a substantial number of issued patents that simply recite "purified" DNA molecules. See for example US Patent Nos. 5,780,262, 6,262,247, 6,399,371, 6,448,042 and 6,555,347. The PTO also routinely issues patents on other purified biomolecules, particularly proteins, based no doubt on the assumption that as a general matter purification of a biomolecule from its native environment is sufficient to confer patent eligibility. See for example US Patent Nos. 6,258,556 and 6,284,236.

In the Federal Circuit's recent decision in AMP v. PTO (discussed briefly already on Patently-O and Patent Docs), the two judges in the majority (Lourie and Moore) held that "isolated" DNA molecules are patent eligible, but implied that mere "purification" of a biomolecule is insufficient to render it patent eligible. Writing for the majority, Judge Lourie correctly observes that patent eligibility under 35 USC 101 of purified biomolecules has never been explicitly addressed by the courts. As I pointed out in an earlier blog post:

The cases most commonly cited for the proposition that a purified naturally occurring compound is patent eligible are In re Kratz, 592 F.2d 1169, 1174 (CCPA 1979) (stating that a naturally occurring strawberry constituent compound does not anticipate claims to the substantially pure compound) and In re Bergstrom, 427 F.2d 1394 (CCPA 1970) (stating that a material occurring in nature in less pure form does not anticipate claims to the pure material). The Federal Circuit implicitly seems to support this view, and as recently as 2003 a Federal Circuit panel cited both Kratz and Bergstrom with apparent approval. But it is worth noting that Kratz and Bergstrom dealt specifically with the novelty and nonobviousness of the compounds, not patent eligibility per se. To my knowledge, there is no judicial precedent that has directly addressed the issue of whether isolation of a naturally occurring molecule renders the isolated molecule and eligible under section 101.

Judge Lourie also correctly noted that the Supreme Court's decision in Funk Bros., although often treated as patent eligibility case, was actually decided on the basis of obviousness, a point that Hal Wegner has long made in which I discussed in a previous blog post.


In what might be the most interesting part of the decision, Judge Lourie makes a clear distinction between purification and isolation of DNA. He explicitly states that "isolated DNA is not purified DNA." In his view, “[p]urification makes pure what was the same material, but was previously impure,” while “[i]solated DNA, in contrast, is a free-standing portion of a native DNA molecule, frequently a single gene. Isolated DNA has been cleaved (i.e., had covalent bonds in its backbone chemically severed) or synthesized to consist of just a fraction of a naturally occurring DNA molecule.”

Judge Lourie stresses that unlike a biomolecule that has been merely purified, isolated DNA "has also been manipulated chemically so as to produce a molecule that is markedly different from that which exists in the body." In his view, the test for patent eligibility hinges on whether the claimed molecule is "markedly different" from that which occurs in nature, and differences at the level of chemical structure between synthetic DNA, or genomic DNA that has been excised from the genome, are sufficient to satisfy the test. Implicitly, he seems to suggest that purified DNA molecules that are not structurally distinct from a naturally occurring counterpart would not be patent eligible. This seems to suggest that claims directed towards purified biomolecules, such as proteins, might not be patent eligible.

In its amicus brief, as I recall, the Biotechnology Industry Organization argued that the District Court's decision in AMP v. PTO should be reversed because it implied that a purified natural product, such as Taxol, would be patent ineligible under the logic of the decision. But in fact it seems that even though the Federal Circuit reversed on the patentability of isolated DNA, its decision suggests that a purified natural product is patent ineligible unless it has distinctions in chemical structure sufficient to render it "markedly different" from its naturally occurring counterpart.

Arguably, the Federal Circuit's decision in AMP v. PTO is not inconsistent with the PTO utility examination guidelines. For example, the guidelines provide two examples of patent eligible isolated DNA - DNA that has been chemically synthesized outside of the body, and DNA that has been excised from the chromosome. Judge Lourie explicitly identifies these two forms of DNA as "isolated" and hence patent eligible.

Judge Lourie and Judge Moore get mixed up at times on the nuances of molecular biology, and make some misstatements regarding the nature of genomic DNA and cDNA. However, while a molecular biologist will pick up on these inaccuracies, I don't think they detract from the core of the decision. In a previous blog post, I explained why the amicus brief filed by the DOJ attempting to distinguish between cDNA and genomic DNA missed the mark because it failed to recognize that, in the vast majority of cases, "isolated" DNA refers to DNA that has been synthesized outside of the native context from which it arose in the body. The DOJ brief (which Judge Moore described as being at times “childlike” in its simplicity) assumed wrongly that "isolated" genomic DNA has merely been plucked from the human cell (what Judge Lourie would characterize as purification), when in fact patents on isolated DNA are based on DNA molecules that have been synthesized in the laboratory, either by cloning into a host cell, or by PCR, or something along those lines. I think that at some level Judge Lourie’s decision gets at this distinction between a biomolecule that has merely been purified, and isolated DNA molecules, which are generally the result of human-directed synthesis.