By Donald Zuhn --
In an article published in
the current issue of Genome Medicine,
Jeffrey Rosenfeld of the University of Medicine & Dentistry of New Jersey
and Christopher Mason of Cornell University contend that due to the
non-specificity of sequence uniqueness across the genome and the broad scope of
claims to nucleotide sequences, the Supreme Court and Congress should limit the
patenting of existing nucleotide sequences ("Pervasive sequence patents cover the entire human genome," Genome Medicine 5:27 (2013)). According to the two researchers, the Association for Molecular Pathology v.
Myriad Genetics, Inc. case affords the Supreme Court with an opportunity to
resolve the "sharp conflict between the public goods of medical knowledge and
improved health and the private goods of rewarding innovation and entrepreneurial
risk-taking" that is presented by gene patents.
The authors begin by stating that "the broadest intellectual property rights on BRCA sequences" come from four claims in Myriad's U.S. Patent No. 5,747,282:
1. An isolated DNA coding for a BRCA1 polypeptide, said polypeptide having the amino acid sequence set forth in SEQ ID NO:2.
2. The isolated DNA of claim 1, wherein said DNA has the nucleotide sequence set forth in SEQ ID NO:1.
5. An isolated DNA having at least 15 nucleotides of the DNA of claim 1.
6. An isolated DNA having at least 15 nucleotides of the DNA of claim 2.
The article takes the Federal Circuit to task for (twice) "declar[ing] that even a short, isolated DNA molecule such as 'ACGT' is different from the 'NNNNN-ACGT-NNNNN' present within a chromosome (AMP v. Myriad, Federal Circuit 2012), because it will not be connected to sugar via a phosphodiester bond and will have a hydroxyl group instead of a bond to a phosphate." According to the authors, the Federal Circuit's ruling means that "even a 15 nucleotide fragment of DNA in Claim #6 from Patent '282 is claimed to be 'markedly different'."
The authors, however, find the Federal Circuit's ruling to be overly broad for three reasons:
First, it relies on the sequences having chemical features and side-chains that are not actually present in the patents . . . : the claims are for a linear series of nucleotides, not a specific chemical structure. Second, if allowed to be so broad, these claims could also create a monopoly on all epigenetic and chemical variations of these sequences. Third, and perhaps most importantly, the non-specificity of 15mer sequences creates unclear infringement liability that has been even noted by the Court.
To investigate this last point, the authors examined the incidence with which 15mers from a given gene matched 15mers in other genes using the Consensus Coding Sequences (CCDS) database of 18,382 high-confidence genes. This analysis showed that "every gene in the CCDS database had a 15mer that matched the sequence of at least one other gene," with the "[t]he number of matching genes ranged from as few as 5 (for MTRNR2L7) or 689 (for BRCA1) to as high as 7,688 (for TTN), corresponding to 0.01%, 4%, and 42% of all genes in the human genome." The authors note that 99.999% of 15mers in the human genome are repeated at least twice.
According to the authors, this analysis "demonstrate[s] that short patent sequences are extremely non-specific and that a 15mer patent claim from one gene will always 'cross-match' and patent a portion of another gene as well." In support of this point, the authors identified 58 patents whose claims covered at least 10% of the bases of all human genes, with the claimed sequences of U.S. Patent No. 7,795,422 matching 91.5% of human genes, and the claimed bovine sequences of U.S. Patent No. 7,468,248 matching 84% of human genes. Arguing that "[t]he demonstrated non-specificity of sequence uniqueness across the genome suggests that the Supreme Court should use [the Myriad] case to clarify the law on gene patents," the authors contend that "[i]f patent claims that use these 15mer or other short k-mer sizes are enforced, it could potentially create a situation where a piece of every gene in the human genome is patented by a phalanx of competing patents, with potentially harmful consequences for genetic testing laboratories and research groups performing targeted sequencing on any gene, in virtually all species."
The authors' conclusion regarding the broad scope of Myriad's 15mer claims, however, is not particularly ground-breaking given that Kepler et al. similarly concluded in a 2010 Genomics paper that claim 5 of the '282 patent was "exceptionally broad" (see "Caught in a Time Warp: The (In)validity of BRCA1 Oligonucleotide Claims"). Kepler et al. suggested that "if human genes were random strings of nucleotides, one would expect a human gene to contain an average of 15 15-mers claimed under the ['282] patent," and in fact found that 80% of 713 human mRNAs deposited in 1994 (the earliest effective filing date of the '282 patent is August 12, 1994) contained at least one of the claimed 15mers. Thus, Kepler et al. indicates that claims 5 and 6 are sufficiently overbroad as to be easily invalidated. Because the analysis of Kepler et al. and Rosenfeld and Mason suggest that claims 5 and 6 of the '282 patent (and perhaps 15mer claims in other patents) would likely not withstand an invalidity challenge, it is unlikely that these claims will have the adverse impact on the "medical good" suggested by the authors of the Genome Medicine article.
One interesting result presented in the Genome Medicine article is that even longer nucleotide fragments from known genes matched the sequence of sequences in other genes. In particular, the paper points out that "even 1,000 nucleotide fragments from known genes could still match 117 other genes." However, the paper does not present any additional information regarding the 1,000 nucleotide fragment (or fragments) that yielded such matches.
Finally, it should be noted that Dr. Mason submitted two declarations when the Myriad case was before the District Court. In a Supplemental Declaration submitted in response to Myriad's brief and Statement of Material Facts, Dr. Mason stated that claim 6 of the '282 patent was "so broad that it includes at least 4% and as much as 100% of the genes in the human genome."
Don,
More malarkey from those that may know the science, but not the patent law that goes with that science. As could be expected, this article even cites the Jensen & Murray study which has been thoroughly discredited by Professor Chris Holman. Also, I'm curious about the data support for the statement in this article that "there are over 40,000 patents on DNA molecules." As Professor Chris Holman has shown with Jensen & Murray study, you need to be careful as to what patents actually cover those "DNA molecules."
Posted by: EG | April 03, 2013 at 09:56 AM
Kevin,
Agree these oligomer claims are likely invalid if challenged. What happens now? Do many patents with "isolated" nucleic acid claims get re-examined? Lawyers simply advise clients not to worry if these are the only claims reducing freedom to operate with a product or service? Litigation? All the above?
Posted by: Bob Cook-Deegan | April 03, 2013 at 10:03 AM
Sorry, last post should have been to Don. Didn't notice it was you at bat today.
BCD
Posted by: Bob Cook-Deegan | April 03, 2013 at 10:06 AM
Bob:
I think your paper (Kepler et al.) and the Rosenfeld and Mason paper indicate that a significant number of 15mer sequences occur more than once in the human genome. For example, Rosenfeld and Mason showed that 15mers from the BRCA1 gene could be found in 689 other genes. To the extent that these other sequences were in the prior art (your paper noted that 80% of 713 human mRNAs deposited in 1994 contained at least one of the claimed BRCA1 15mers), this could create an issue with respect to the novelty of the 15mer claims (i.e., claims 5 and 6 of the '282 patent).
Thank you for the comments,
Don
Posted by: Donald Zuhn | April 03, 2013 at 10:23 AM
Don,
The previous work made an estimate based on cDNA; we wanted to expand this and look at the actual empirical distribution of k-mers as a patent - all of which we show to be non-specific up to 1000mers.
Also, we used updated patents of DNA that are actually claimed in the language of the claims, specifically to address the Holman article. But, even if you take out 99% of those, you still have enough patents which do claim the sequences that easily match 91% of human genes. It's just not that hard to do with short fragments, and Venter tried this back in the 90s as well. While you may bemoan scientists dipping into the law, we only do it because the laws (in this case, the patents) can sometimes blatantly disregard the science, which only makes our jobs harder. Thanks for the debate.
Posted by: Christopher Mason | April 04, 2013 at 07:04 AM
Chris,
In your paper you provide examples of claims that have roughly the following format:
1. An isolated DNA having the nucleotide sequence [5,000 specific bp]
2. An isolated DNA having at least 15 nucleotides of the DNA of claim 1.
Your problem seems to relate to the second claim, not the first, is that correct?
Posted by: Moocow | April 04, 2013 at 12:52 PM
There's quite a few Ph.D. patent attorneys on this blog, no? It seems odd that they are unwilling to simply come out and say whether any of Myriad's claims are anticipated or not. I can't tell if they are genuinely unconvinced, if they are confused, if they are feigning confusion, if they are simply hiding from facts, or if they are afraid to take a stand.
What's the problem, guys?
Posted by: Bryan Rudner | April 09, 2013 at 11:23 AM