By Professor Christopher Holman
[The following is a blog posting by Christopher Holman, a patent law scholar at UMKC School of Law, that he originally posted on April 5, 2013 at his blog, Holman’s Biotech IP Blog, where Professor Holman regularly blogs on important issues in biotech and IP law. Professor Holman kindly gave us permission to repost his blog posting here.]
A Critique of a Recent Article Which Found That Sequence Patents Cover the Entire Human Genome
By Christopher Holman
I recently fielded a phone call from a reporter with a leading international scientific journal, asking for my opinion of an article entitled “Pervasive Sequence Patents Cover the Entire Human Genome,”recently published in a publication called Genome Medicine. I have published several articles debunking the myth that 20% of human genes are patented, and the reporter thought that the article in Genome Medicine, authored by a researcher affiliated with Yale Law School’s The Information Society Project, contradicted the results of my study. I took a look at the “Pervasive Sequence Patents” article and found it to be a fundamentally flawed empirical study that will sadly be used to further support the widespread misperception that access to a large percentage of the human genome is precluded by a thicket of gene patents.
The “Pervasive Sequence Patents” article does cite to my 2012 Nature Biotechnology article Debunking the Myth That Whole Genome Sequencing Infringes Thousands of Gene Patents, but the authors apparently missed the main point I was trying to make. The myth that 20% of human genes are patented was born out of a 2005 article published in Science by Jensen and Murray that found that the sequence of 20% of human genes (or in some cases the protein encoded by human gene) is mentioned in a US patent claim. The problem arose when people assumed that the mention of a gene’s DNA sequence in a patent claim is equivalent to the patenting of the gene, which led to an assumption that any use of or research on any of these genes would result in patent infringement. In my article, I explained that in patent law “the name of the game is the claim “(to quote Judge Rich), and that when one actually reads the patent claims in the patents identified by Jensen and Murray it is clear that few if any of the patents would be infringed by many forms of research or genetic testing, including diagnostic testing and whole genome sequencing.
Unfortunately, the authors of “Pervasive Sequence Patents” have apparently fallen into the same trap, assuming that mention of a gene’s DNA sequence in a patent claim results in the patenting of the gene in a manner that totally blocks access to the gene. Even more problematically, the authors seem to assume that every patent with a claim mentioning a gene sequence also claims every 15mer present in the sequence, i.e., every contiguous 15 nucleotide sequence appearing in the gene. Presumably they made this assumption because the Myriad gene patent litigation includes a patent claim directed to 15mers of the BRCA1 encoding sequence, including Claim 5 from US patent number 5,747,282:
An isolated DNA having at least 15 nucleotides of the DNA of claim 1.
There are two fundamental problems with this empirical approach. One is that it does not necessarily follow that the mention of a gene’s DNA sequence in a claim equates with the patenting of the gene – that was the main point of my Nature Biotechnology article. The other is to assume that all of these patents include claims analogous to Claim 5 of the ‘282 patent.
In my experience, claims of this type are extremely rare. I looked at hundred patents identified as gene patents in the Jensen Murray study and found that most only claim the full-length gene sequence, and if fragments were claimed the fragments are much larger than 15 nucleotides. In fact, I looked through hundreds of gene patents trying to find another 15mer claim analogous to those in the Myriad patents and could not find one. The patent claims at issue in the Myriad case will be expiring within the next few years I believe, and I doubt that this sort of broad 15mer claim has been issued by the patent office in recent years, or if it has it seems to be extremely rare.
In any event, in 2010 Keppler et al. published an article entitled “Metastasizing Patent Claims in BRCA1” which showed that if the BRCA 15mer claims are interpreted so broadly as to cover any DNA sequence comprising any 15mer appearing in a BRCA gene, there appears to be a wealth of prior art that would invalidate the claim regardless of the claims patent eligibility.
The flawed methodology used in the “Pervasive Sequence Patents” article is readily apparent from the results of their empirical study. Here is what they reported as the result of their study:
[W]hen we took existing gene patents and matched their 15mers to known genes, we found that 100% of known genes have at least one 15mer claimed in a known patent. Current gene patents were observed to match each gene many times, with 1,295 matches to other genes on average (standard deviation 1,208). When we examined the amount of total sequence space in human genes that is covered by 15mers in claims from current patents (Additional file 2), we found 58 patents whose claims covered at least 10% of the bases of all human genes. The top patent was US7795422, whose claims’ sequences matched 91.5% of human genes. Interestingly, we also observed a patent for improving bovine traits (US7468248) with explicit claims for 15mers that matched 84% of human genes. This patent was not even aimed at any human sequence, yet covered a majority of human genes once we examined the claim’s matches at the 15mer scale.
First off, let’s look at the “top patent” they found, US7795422, “whose claims sequences matched 91.5% of human genes.” The ‘422 patent has only one independent claim:
1. A chemically modified short interfering nucleic acid (siNA) molecule, wherein: (a) the siNA molecule comprises a sense strand and an antisense strand, each strand having one or more pyrimidine nucleotides and one or more purine nucleotides; (b) each strand is independently 18 to 27 nucleotides in length, and together comprise a duplex having between 17 and 23 base pairs; (c) the antisense strand is complementary to a human Hypoxia Inducible Factor 1 (HIF1) RNA sequence comprising SEQ ID NO:567; (d) a plurality of pyrimidine nucleotides present in the sense strand are 2′-deoxy-2-fluoro pyrimidine nucleotides and a plurality of purine nucleotide present in the sense strand are 2′-deoxy purine nucleotides; and (e) a plurality of pyrimidine nucleotides present in the antisense strand are 2′-deoxy-2′-fluoro pyrimidine nucleotides and a plurality of purine nucleotides present in the antisense strand are 2′-O-methyl-puine nucleotides.
When one reads the claim, it is apparent on the face that the claim is limited to “chemically modified” molecules comprising 2′-deoxy-2-fluoro pyrimidine nucleotides and 2′-deoxy purine nucleotides. DNA does not contain 2′-deoxy-2-fluoro pyrimidine nucleotides and 2′-deoxy purine nucleotides, these are synthetic analogues to the nucleotides that appear in DNA. This patent that the authors found to match 91.5% of human genes does not cover any gene or any DNA molecule, only chemically modified synthetic molecules for use in RNA interference.
Next the authors reported that US7468248 contains “explicit claims for 15mers that matched 84% of human genes.” In fact, the ‘248 patent has only two independent claims, both of them method claims:
1. A method for inferring a trait of a bovine subject from a nucleic acid sample of the bovine subject, comprising identifying in the nucleic acid sample, a nucleotide occurrence of a single nucleotide polymorphism (SNP) at position 300 of SEQ ID NO:21645, thereby inferring the trait, wherein the trait is marbling, tenderness, fat thickness, red meat yield, or average daily weight gain.
22. A method for determining a nucleotide occurrence of a polymorphism in a bovine sample, comprising: a) contacting a bovine polynucleotide in the sample with an oligonucleotide that binds to a target region, wherein the target region comprises a position at position 300 of SEQ ID NO:21645 or wherein the target region is within 3000 nucleotides of a nucleotide at position 300 of SEQ ID NO:21645, and b) determining the nucleotide occurrence of a single nucleotide polymorphism (SNP) at position 300 of SEQ ID NO:21645, wherein the determination comprises analyzing binding of the oligonucleotide or detecting an amplification product generated using the oligonucleotide, thereby determining the nucleotide occurrence of the polymorphism.
Both of these claims would only be infringed by someone performing a specific genetic test on a bovine subject (colloquially a cow). The patent does not include any claim covering any DNA sequence, and the authors’ assumption that the patent “explicitly claims 15mers that matched 84% of human genes” implies that they either did not read the claims or do not understand the basics of claim interpretation.
The problems with this article are pretty apparent once one reads the claims of the patents that were identified as “matching” human genes. Unfortunately, it is just the latest installment of a prolific stream of fundamentally flawed academic articles that are being cited in support of the notion that human gene patents are a pervasive problem. I don’t doubt that the authors meant well, but it’s dangerous to conduct empirical patent studies without appreciating and understanding the critical role of the patent claim. And the publication of the article highlights the limitations of peer review (assuming Genome Medicine engages in peer review).