Pervasive sequence patents cover the entire human genome
© BioMed Central Ltd 2013
Published: 25 March 2013
Skip to main content
© BioMed Central Ltd 2013
Published: 25 March 2013
The scope and eligibility of patents for genetic sequences have been debated for decades, but a critical case regarding gene patents (Association of Molecular Pathologists v. Myriad Genetics) is now reaching the US Supreme Court. Recent court rulings have supported the assertion that such patents can provide intellectual property rights on sequences as small as 15 nucleotides (15mers), but an analysis of all current US patent claims and the human genome presented here shows that 15mer sequences from all human genes match at least one other gene. The average gene matches 364 other genes as 15mers; the breast-cancer-associated gene BRCA1 has 15mers matching at least 689 other genes. Longer sequences (1,000 bp) still showed extensive cross-gene matches. Furthermore, 15mer-length claims from bovine and other animal patents could also claim as much as 84% of the genes in the human genome. In addition, when we expanded our analysis to full-length patent claims on DNA from all US patents to date, we found that 41% of the genes in the human genome have been claimed. Thus, current patents for both short and long nucleotide sequences are extraordinarily non-specific and create an uncertain, problematic liability for genomic medicine, especially in regard to targeted re-sequencing and other sequence diagnostic assays.
Some DNA patents are for a very specific series of nucleotides (such as 5'-ATGCGACGGATCGATC-3') or an exact chemical structure (such as a DNA molecule modified with a fluorescent probe), but diagnostic DNA-based patents have broader claims . These patents are used to find mutations in various disease-related genes, and the specified DNA sequence as well as any other similar sequence are often covered within the patent claim. This is because there are many (at least [(2N) - 1]) possible combinations of mutations for a gene . Diagnostic gene patents are therefore written to find any known or unknown variation of a gene. For example, in the Association of Molecular Pathologists (AMP) v. Myriad case, the broadest intellectual property rights on BRCA sequences come from several related claims in patent 5,747,282:
Claim #1. An isolated DNA coding for a BRCA1 polypeptide, said polypeptide having the amino acid sequence set forth in SEQ ID NO:2 (the BRCA1 cDNA).
Claim #2. The isolated DNA of claim 1, wherein said DNA has the nucleotide sequence set forth in SEQ ID NO:1 (the BRCA1 gene).
Claim #5. An isolated DNA having at least 15 nucleotides of the DNA of claim 1
Claim #6. An isolated DNA having at least 15 nucleotides of the DNA of claim 2.
In 2010, in the first ruling on AMP v. Myriad, a US District Court stated that all of these patent claims were invalid and that isolated DNA is 'not patentable subject matter'. Then, in 2011, the US Court of Appeals for the Federal Circuit overturned this ruling (2 to 1 decision) and stated that an isolated DNA molecule is 'markedly different' from native genomic DNA and that fragments of the BRCA genes can be patented. After a re-hearing in light of another case (Mayo v. Prometheus), the same decision was issued by the Federal Circuit in August 2012. Recently, the Supreme Court decided to hear arguments in the case, opening a re-examination of the lower courts' decisions and rationale on these gene patents and their claims.
Establishing the uniqueness of gene patents in DNA and cDNA could potentially have a large impact on the interpretation of these patents [7, 8]. Previous work has examined Claim #5 with respect to other cDNAs on one chromosome (Chr1), and used these data to estimate that 15 infringing 15mers might exist for any cDNA . However, these studies compared the likely uniqueness of cDNAs on the basis of an average degeneracy of the genetic code, leaving open the issue of exact DNA matches in the coding regions of genes and the genome. Also, a first estimate, made in 2005, calculated that 18% of human genes were patented , but many new DNA patents have emerged since (Figure 1). These results have been challenged in recent work , which has demonstrated that some gene patents for genetic sequences do not contain the DNA fragments within the actual claims. Thus, we sought to examine the current landscape of gene patents using empirical, exact matches to known genes that were confirmed to be present in patent claims, ranging from sequences of 15 nucleotides (15mers) to the full lengths of all patented DNA fragments.
We then examined the distribution of 'cross-matches' for varying k-mers across all human genes. We found that the number of matches decreased as the k-mer size increased, showing an inversely proportional relationship of sequence uniqueness and k-mer size (Figure 3b). Notably, even 1,000 nucleotide fragments from known genes could still match 117 other genes, showing that long gene fragments can still show substantial non-specificity. We then used the same alignment criteria to examine the uniqueness of the entire human genome (beyond just coding regions), and we found that 99.999% of 15mers in the human genome are repeated at least twice (see Methods and data). These data confirm the findings of previous studies that showed little sequence specificity for small k-mers in the human genome [13–15], but our data show for the first time that this global non-specificity of 15mers and longer k-mers impacts all gene patents, including those on BRCA1 and those claiming non-coding areas of the genome.
Additionally, when 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.
These results have striking implications for the AMP v. Myriad Genetics case, gene patent litigation, and other patent legislation. The demonstrated non-specificity of sequence uniqueness across the genome suggests that the Supreme Court should use this case to clarify the law on gene patents. If 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.
Our data show that currently claimed nucleotide sequences in US patents cover at least 41% of existing genes, as identified by BLAST alignment matches, and as many as 100% when allowing for 15mers. We also observed a large number of cross-kingdom exact matches of 15mers, indicating that not only human genetic sequences are in play: entire patent families from plant genomics can similarly claim the majority of the human genome. As both plant and animal patents claim short sequences, as well as those with low homology (45 to 55%), any claimed sequence will inevitably match many others. Yet, importantly, we observe that there is no ideal k-mer size that will preclude matches with another gene in the genome. Thus, the non-specificity needed for diagnostic patents to find any mutated sequence of one gene expands their property rights to hundreds or thousands of other genes.
Some have commented that whole-genome sequencing (WGS) might avoid some of these infringement problems , as no targeted fragment that contains a patented sequence is specifically isolated when shotgun-based approaches are used for sequencing. There is, however, no specific case precedent that yet supports this conclusion, and as such, legal ambiguities still affect WGS. These infringement risks are also still very clear for PCR-based or enrichment-capture strategies, which directly overlap with these patents by enriching for a molecule that contains BRCA1 or the targets of other patented genes. These targeted methods currently represent the vast majority of clinical sequencing for diagnostic medical decisions in molecular pathology, and they will likely be used for a long time as validation technology for any WGS approaches.
Fortunately, there is precedence in US history for resolving such a medical-legal conflict. In 1992, a US patent was issued (#5,080,111) for a 'self-sealing episcleral incision', and this patent required a license of $4.00 per operation for any doctor to use the method. A lawsuit was filed against the patent by the American Medical Association (AMA), which strongly condemned 'the patenting of medical and surgical procedures' and began to work with Congress to outlaw the practice. The AMA argued that the best surgical techniques should not be denied to patients, simply because of legal reasons or fees. In 1995, the United States Code was changed accordingly, to add language that exempted any 'patient, physician, licensed healthcare practitioner, or a health care entity from infringement of a patented medical or surgical procedure, therapy, or diagnosis'. Thus, any surgeon could still patent a new method and stake claim as the inventor, but the patients' need to get access to the best medical care out-weighed the infringement of the intellectual property rights of surgeons.
Now that the era of genomic medicine is here, the US Supreme Court has the chance to shape the balance of the medical good and inventor protection. Given the falling price of genome sequencing and targeted re-sequencing, and the ubiquity of the genomics technologies, the urgency to resolve this uncertainly around gene patents has never been more salient. Failure to resolve these ambiguities perpetuates a direct threat to 'genomic liberty', or the right to examine one's own DNA. Our analysis and data provide strong evidence that the Supreme Court and Congress should limit the patenting of existing nucleotide sequences because of their broad scope and non-specificity in the human genome. Finally, we suggest that a research exemption or limited liability for patent infringement be implemented, as has been done for surgical techniques, which could craft a functional balance between the rights of inventors and the best application of personalized genomics for improved medical care.
We used gene sequences from the CCDS database, which were downloaded from the NCBI web site . This is a curated set of high-confidence genes prepared by an international consortium . It contains 18,382 genes having a total of 26,355 isoforms. For this analysis, we used the longest isoform of each gene as a single, representative sequence for that gene.
We used the public database from The Cancer Genome Atlas (TCGA) , and specifically the query and search tool at Memorial Sloan Kettering Cancer Center (MSKCC) . We used the query 'BRCA1' for the database search. Data queried on 9 March 2013.
Patents in the inClaims subset of the Cambia  patent sequence database were downloaded from . This database has been curated to include all of the sequences from US Patents that are specifically contained within the claims of the inventions, and we validated 25 of these manually (from nt-inClaims.fsa) to confirm that they were correct.
For the comparisons of k-mers between genes, we used Bowtie (v0.12.5) to report all of the matches for each query. Each CCDS gene was split into all overlapping k-mers of the designated length and these sequences were searched against all of the other genes.
The comparison of patents to genes was performed using the criteria used by Jensen and Murray ; specifically using BLAST version 2.2.25+  to match at least 150 nucleotides and an e-value of 0. This was performed using the BLAST command: blastn -evalue 1e-307 -word_size 150.
In addition, the criteria of e-value <0.05 was used, which is a typical set of parameters for BLAST. The command was: blastn -evalue 0.05. For the exact matching of 15 bp sequences, Bowtie was used. The bowtie command was: bowtie -f -a -best -v 0.
For the calculations in our text about mutational complexity, we show that for a gene with N mutations, there are [(2N) - 1] combinations of mutated forms of that gene. For example, given a sequence with ten bases, where '=' represents an un-modified base pair, where three mutations (A,G,T) exist, there are [(2N) - 1] = [23 - 1] = [8 - 1] = 7 possible combinations.
These seven possible mutated forms that are different from the wild-type are shown here:
The results published here are in whole or part based upon data generated by The Cancer Genome Atlas pilot project established by the NCI and NHGRI. Information about TCGA and the investigators and institutions who constitute the TCGA research network can be found at the project website .
American Medical Association
Association of Molecular Pathologists
Basic Local Alignment Search Tool
Consensus Coding Sequences