Faced with a growing pool of genomic data and the rising complexity of biotech patent applications, the US Patent and Trademark Office responded earlier this year by restricting the number of sequences that can be included in a single patent application.
In an Official Gazette notice in February, the USPTO rescinded a 1996 policy that allowed patent applicants to claim up to 10 independent nucleotide sequences in a single application.
In the notice, the agency said that because of the “exponential” increase in genomic databases over the past decade, “it now requires significantly more computational time to run individual nucleotide sequence searches for examination purposes than in 1996, and there is significantly more pertinent prior art to consider.”
The USPTO said the additional time for sequence searching “consumes a disproportionate amount of office resources over that expended in 1996,” and drove the decision to impose the new limit.
A USPTO spokesperson stressed that the notice is not an across-the-board prohibition on applications with more than one sequence, but rather an attempt to leave that decision up to patent examiners. The ruling states that the USPTO “will no longer waive its restriction process with respect to nucleic acid sequences and allow a reasonable number, up to 10, to be searched at one time, unless it would be reasonable to do,” the spokesperson said.
“As always, patent examiners have discretion to examine as many nucleic acid sequences as they see fit and that [are] reasonable to submit for search based on computational resources and [to] examine in one application.”
The ruling could have a significant impact on biotech and pharmaceutical companies, who must now file separate applications — and pay the subsequent filing, search, and examination fees — for each sequence.
The 1996 policy that the USPTO rescinded had waived a standing rule that patent applications must relate to a single claimed invention, in this case a single nucleotide sequence.
“The notice now places biotechnology applications claiming nucleic acid sequences on the same standing as all other applications with respect to whether an application relates to more than a single invention,” a USPTO spokesperson told BioInform via e-mail.
According to bioinformatics firm GenomeQuest, which specializes in sequence search software for intellectual property professionals, around two-thirds of sequence-based patent filings currently include more than 10 sequences.
In an analysis of its internal database, GenomeQuest found that patents with between one and 10 sequences made up the largest portion of filings each year between 2001 and the first six months of 2007, but that portion has actually declined from 38 percent of all filings in 2001 to 33.33 percent of all filings so far in 2007 (see accompanying chart).
Michael McManus, vice president and general manager of GenomeQuest, said it’s too early to see the effect of the February ruling among his company’s clientele. He told BioInform that even though he hasn’t had direct feedback from clients regarding the limits, “If a company is used to filing thousands of sequences” and they have to now limit that to one, “it seems obvious” that it would be inconvenient for them.
Timothy McBride, an attorney with Senniger Powers in St. Louis, Mo., explained via e-mail that the effect of the notice is that “claims to polynucleotide molecules are to be ‘considered for independence, relatedness, distinction and burden’ just as are claims to any other type of invention, and the ‘special treatment’ that nucleotide sequences had previously received was rescinded.”
More significantly, he said, “the USPTO can now restrict the claims of an application such that only a single polynucleotide molecule is examined per application.”
Is it a Bioinformatics Problem?
When the USPTO first issued the 1996 waiver that permitted examiners to review up to 10 sequences in a single application, genome sequencing was in its infancy, and most sequence-based applications included cDNAs discovered by expressed sequence tag techniques.
Writing in the Official Gazette notice, the USPTO said “At that time, and in many of those applications, little information was provided relating to function of the nucleic acid, nor was there significant description of the function or the information content (e.g., protein coding capacity) of the nucleic acid claimed.” “Consequently such claims were, in many instances, simple in format and narrow in scope.”
Since then, sequencing has become commonplace, causing GenBank to experience a ninety-fold increase in the number of nucleotides and about a fifty-four-fold increase in the number of sequences.
In addition, the nomenclature surrounding polynucleotide molecules has become more complex, the USPTO said. “Polynucleotide molecules are now often claimed in a single application in a variety of complex formats, some of which may embrace multiple inventions, such as by reference to: the amino acid sequence of the protein encoded; the ATCC number of a deposited plasmid containing the polynucleotide molecule; arbitrary laboratory designations; function of the nucleic acid alone or in combination with a partial linear nucleotide sequence; a genus described in terms of homology, percent identity, or hybridization; a genus (or subgenus) described by nucleic acid sequence with variable positions specified within the sequence listing; SNPs; antisense; or interfering RNA.”
In addition to the computational resources required to run a multi-sequence search, the agency said, “it currently takes more office resources to correlate the claimed polynucleotide with the polynucleotide as defined in the prior art because it is increasingly common for both patent applicants and prior art references to describe a polynucleotide molecule in different ways.”
In a statement submitted to BioInform, the USPTO said that the agency’s IT efforts “have kept pace with the increasing demands for examination of applications, in accordance with relevant law and regulation.” The USPTO is obligated to “meet the search and examination burden for one invention per application; implicit in this is that two or more structurally and functionally unrelated sequences constitute two or more unrelated inventions,” the agency said.
Compounding the problem is that the agency runs the highly sensitive but computationally demanding Smith-Waterman sequence-search algorithm rather than heuristic approaches like Blast.
“It is not just a computational challenge; it is also an examiner challenge to reviewing the ever-increasing amount of search result output.”
“The sensitivity of this algorithm is necessary for proper identification of relevant prior art,” the agency said. “Faster but less sensitive heuristic methods, such as Blast, meet many needs of researchers, but are not sufficiently sensitive for a thorough identification of prior art”.
The USPTO has developed a range of internal tools for managing and analyzing sequence data — including ABSS, the Automated Biotechnology Sequence Search system; SCORE, the Supplemental Complex Repository for Examiners; PSIPS, the Publication Site for Issued and Published Squences; and PatentIn, an authoring tool for patent applicants to use when submitting sequences (see sidebar for further details on these tools).
The USPTO said in its statement that it has been “continually addressing the sequence search challenges by upgrading computer systems regularly and seeking improvement of the resources.”
However, the agency said, “The problem is that the human review of the output is now far more time-consuming. Technical advances go only so far, since the amount of output examiners need to review has grown. It is not just a computational challenge; it is also an examiner challenge to reviewing the ever-increasing amount of search result output.”
The agency faces further resource constraints. While the complexity of these patent applications is increasing, the overall number of applications has remained stable over the last several years, so hiring additional examiners to keep pace with the growth in genomics data is not an option.
“Technology Center 1600, which examines patent applications drawn to nucleotide sequences, has hired and currently has on board a maximum number of patent examiners to address the current workload of applications of this nature,” the agency told BioInform. “Filings in this area have not reflected a substantial growth in recent times and thus, it is believed that the agency has sufficient number of examiners in these areas.”
More Filings, More Fees
Senniger Powers’ McBride agreed that the USPTO is “understaffed and overworked. Therefore, the reduced burden of searching and examining a single nucleotide sequence per application as opposed to ‘up to ten’ sequences per application certainly plays into the issuance of the 2007 notice.”
However, he noted that the ruling may also be related to a broader effort by the USPTO to increase revenues.
“Within the past year or two, the USPTO has been much more aggressive with its restriction practice (regardless of the claimed subject matter), issuing restriction requirements that dice patent application claim sets into numerous groups of claims asserted by the office to be ‘independent and distinct inventions.’”
He said that the result is that patent applicants are forced to divide their single patent application into multiple divisional applications, paying government filing fees for each additional application. “As a result, the USPTO receives multiple government filing fees for the examination of applications that would otherwise have been examined as a single application with the payment of a single filing fee.”
— Laurie Wiegler contributed to this article.