Influenza Genome Executive Summary

The possibility of pandemic influenza arising from the avian influenza virus, H5N1, is of critical importance for global public health.   The need to safeguard the health of the poor and to provide fair access to needed medicines, diagnostics, and vaccines has put patent systems at the heart of the debate.   How do patents impact development and commercialization of products targeted against major world diseases?   Does the use of patents as a part of strategy to inventivize commercial development adversely and disproportionately affect some nations and populations?  Do internationally shared public health goals - such as addressing pandemic influenza - require different treatment of patents than public health challenges that are regional in nature?   Virtually all of the discussion of this topic occurs in the absence of a clear and transparent body of evidence.  

The most strongly voiced concern is whether patents claiming influenza genomes, genes or proteins will compromise equitable access to products that are derived from them.    It is not surprising that the nations most strongly affected by avian flu are most deeply concerned about their ability to respond to current and projected public health needs.   Global health officials and national authorities seek clarity in the terms of access to health interventions to address pandemic influenza.   To address this concern, a factual basis for how much of the influenza genome, especially of the H5N1 virus and its derivatives, has been patented or has patents pending is needed.

If the extent of influenza genome-related patents can be determined, and their claims and impacts can be analyzed, then their potential effect can be meaningfully discussed and modeled.    To do this type of comprehensive analysis of genome-related patents will require a many avenues of investigation, new tools and diverse approaches.  The results of one component of this analysis - identification and interpretation of influenza sequences in claims of United States patents and patent applications - is presented in this landscape.   It is part of a wider study initiated by World Intellectual Property Organization (WIPO), and has been funded in part by PATH Vaccine Solution.  Other complementary approaches are reported on the WIPO web site.    Using the tools developed in this study, the next stages will examine critical jurisdictions whose patent systems are much less navigable.

The patent landscape determined by sequence analysis has yielded some key findings:

Claims to nucleotide sequences -

• One-half of the patents are directed to equine influenza virus (assignee = Heska Corp.);
• None appear to claim the H5 subtype of segment 4, which encodes hemagglutinin (HA) or the N1 subtype of segment 6, which encodes neuraminidase (NA);
• Of the two patents that claim HA sequences, one claims sequence encoding the signal peptide region from HA1 and HA3 (US 6245532), and the other claims an antigenic fragment of HA2 and HA3;

• Claims of three patent applications are directed to reassortment viruses, useful as vaccines against H5N1 (US 20060008473 - MedImmune) or H1N1 (US 20050003349 and US 20070231348 - Wisconsin Alumni Research Foundation);
• Of the remaining patent applications that initially claimed influenza nucleic acids, only about one-half still claim nucleic acid sequences and most of these are directed to sequences encoding internal proteins.
• For the granted patents with nucleotide claims, 75% were assigned to corporations and 25% were assigned to non-profit organizations. For the patent applications, 47% were assigned to corporations and 53% were assigned to non-profit organizations.

Claims to amino acid sequences -

• Most of the patent claims are directed to short peptides;

• Most of the patent claims reciting full-length or near full-length proteins are drawn to equine influenza virus proteins;
• Patent claims can be open-ended with respect to sub-types of influenza proteins, such as claiming a reassortant virus in which the sequences of the six internal proteins are specified but the sequences of hemagglutinin (HA) and neuraminidase (NA) are unspecified;

• Peptides in the claims may be used for a variety of purposes, and especially for vaccines;
• Claims of the patent applications are similar to those from patent grants, except that the number of sequences in the claims is higher.
• For the granted patents with peptide claims, 45% were assigned to corporations and 55% were assigned to non-profit organizations. For the patent applications, 70% were assigned to corporations and 30% were assigned to non-profit organizations.

Limitations of landscape analysis

In doing this study, it has become clear that we face a serious limitation with the virtual impossibility of searching patents in many of the countries currently impacted by Avian Influenza. These Asian countries are also potential hot-spots for the evolution of a pandemic influenza strain. Some of them have expressed profound concerns about the extent of patenting over inventions derived from strains they may contribute to the international influenza efforts. In general, however, the patent offices in South, South East and Eastern Asia, do not have text searchable patents. 

Of necessity therefore, this study has been restricted to jurisdictions which have such searching capabilities or report filings of patent applications to international organizations, such as the European Patent Office, that disseminate the data. In so doing, we hope to at least see which players with international ambitions may be pursuing patent protection over influenza genome - related inventions, to better focus particular targeted searches through this means.

These countries are in a key position to create a basis of evidence from which to evaluate the merits of these concerns and to develop remedies should these concerns be validated in their countries.  In furtherance of this aim, we have been extending the Patent Lens search and analysis capabilities to internationalize and make ready to incorporate data from these critical jurisdictions as they become available - and extend the methodology of this first draft landscape to the new data.

Even for those jurisdictions with some searching facility, finding DNA or protein sequences disclosed in or claimed in patents is extraordinarily difficult if not impossible.  We also appreciate that there are many techniques for filing invention disclosures that render the searchability of DNA or protein sequences very difficult, including filing PCT applications with 'image-based' sequence files that require optical character recognition (OCR). Thus, this study has stimulated an accelerated program within Patent Lens for improved OCR of sequence files from PCT applications.

Photo Caption:  A negative-stained transmission electron micrograph (TEM) depicting the ultrastructural details of an influenza virus particle. Photo courtesy of the CDC.

Background

The possibility of pandemic influenza arising from the avian influenza virus, H5N1, is of critical importance for global public health.   The need to safeguard the health of the poor and to provide fair access to needed medicines, diagnostics, and vaccines has put patent systems at the heart of the debate.   How do patents impact development and commercialization of products targeted against major world diseases?   Does the use of patents as a part of strategy to incentivize commercial development adversely and disproportionately affect some nations and populations?  Do internationally shared public health goals - such as addressing pandemic influenza - require different treatment of patents than public health challenges that are regional in nature?   Virtually all of the discussion of this topic occurs in the absence of a clear and transparent body of evidence.  

The most strongly voiced concern is whether patents claiming influenza genomes, genes or proteins will compromise equitable access to products that are derived from them.    It is not surprising that the nations most strongly affected by avian flu are most deeply concerned about their ability to respond to current and projected public health needs.   Global health officials and national authorities seek clarity in the terms of access to health interventions to address pandemic influenza.   To address this concern, a factual basis for how much of the influenza genome, especially of the H5N1 virus and its derivatives, has been patented or has patents pending is needed.

If the extent of influenza genome-related patents can be determined, and their claims and impacts can be analyzed, then their potential effect can be meaningfully discussed and modeled.    To do this type of comprehensive analysis of genome-related patents will require a many avenues of investigation, new tools and diverse approaches.  The results of one component of this analysis - identification and interpretation of influenza sequences in claims of United States patents and patent applications - is presented in this landscape.   It is part of a wider study initiated by World Intellectual Property Organization (WIPO), and has been funded in part by PATH Vaccine Solution.  Other complementary approaches are reported on the WIPO web site.    Using the tools developed in this study, the next stages will examine critical jurisdictions whose patent systems are much less navigable.

Are influenza genomes recited in patent claims?  And if so, what portions of which genomes? Are genes from the influenza genome claimed in patents or applications?   Are proteins or variants from these genes claimed?

In the first phase of this study, the patents disclosing influenza sequences, and their claims were found by a search of the sequences themselves, rather than by a search using keywords.   Our approach incorporated sequence alignment tools to compare a representative sample of influenza nucleotide and amino acid sequences to collections of nucleotide and amino acid sequences that are specifically recited in the claims of granted patents and pending patent applications in the United States. A sequence-based approach may find patents and patent applications that aren't found in conventional keyword-based searches.  In the effort to to perform a comprehensive patent landscape, the two approaches should be seen as complementary.

Photo caption: A Thai worker sprays disinfectant on chicken cages Tuesday, July 12, 2005, in the central province of Suphanburi, Thailand, 100 kilometers (60 miles) north of Bangkok. Photo courtesy of the U.S. Department of State (2005: The Year in Pictures-Strengthening Alliances and Defending Principles).

Influenza Infection

Influenza is a contagious disease of mainly the upper respiratory tract (nose, throat, bronchi) and is caused by a virus.  The virus is passed from person to person by viral particles present in aerosols generated by sneezes and coughs.  In most people, the infection is short-lived, about one to two weeks, and typically people recover on their own without complications.  In the very young, elderly and people having other diseases, influenza can take a more severe course and lead to complications and even death.  For such people, yearly vaccination is indicated. Influenza rapidly spreads around the world in seasonal epidemics, with new strains arising in East or Southeast Asia, traveling to Australia and other parts of Oceania, to Western Asia and Europe, to North America, and finally to South America where the strain dies out (Science, 2008).  Each year a new strain or strains become the infectious agent.  Different strains represent the result of minor genetic changes in the influenza gene sequences. 

Two different influenza viruses cause human disease:  influenza A and influenza B.  Three subtypes of influenza A are important for humans:  A(H3N2), A(H1N1) and A(H1N2), of which A(H3N2) is currently associated with the most deaths. The H and N are shorthand for two proteins, haemagglutinin (H) and neuraminidase (N), located on the surface of the virus.  Most recently, there have been limited outbreaks of A(H5N1) virus, which has been transmitted from birds to humans.  The H5N1 virus is potent, it is highly contagious in birds and can be deadly to them, moreover, over half of the people that contracted this influenza have died.  Most of these people have been in direct or close contact with infected poultry. In general, H5N1 remains a very rare disease in people and does not appear to readily spread from person to person.  

Nonetheless, because all influenza viruses have the ability to change, public health officials and scientists are concerned that A(H5N1) virus could change to become infectious and contagious in humans.  If A(H5N1) virus were to gain the capacity to spread easily from person to person, there is a likelihood of an influenza pandemic (worldwide outbreak of disease) could begin. Based on statistics from past influenza pandemics, and the known morbidity of the few cases of human A(H5N1) disease, a high death rate and a very large number of deaths could occur.  To limit these disastrous consequences, specific vaccines are being developed. Vaccination is the principal way of attacking influenza, for many reasons antiviral medications are not generally used.  

For more information on influenza, H5N1 virus and pandemics, see

CDC - Center for Disease Control
WHO - World Health Organization

Photo Caption:  Policemen in Seattle wearing masks made by the Red Cross during the influenza epidemic in December, 1918. Photo courtesy of the National Archives and Records Administration.

The Influenza Genome Comprises Eight Segments

In contrast to many other viruses, the influenza genome is composed of RNA rather than DNA. The genome comprises eight segments, and only virus particles containing all eight segments are viable. New genomes can be assembled as a result of mixing between two viruses. That is, if a cell is infected with two different strains of influenza, it is possible for the segments to re-assort to create new viruses that have segments from each original virus. New combinations may be more or less virulent than the old combinations. Figure 1 below shows the relative sizes of the eight influenza segments as well as the genes that are specified by each. Figure 2 shows how the segments are arranged during viral replication.

Figure 1: The eight RNA segments of the influenza genome.

Figure 2: The systematic arrangement of genes during flu virus replication.

(Figure 2 courtesy of the National Institute of Allergy and Infectious Diseases)

Influenza Genome Resources

There are a number of resources that provide information about influenza, but perhaps the most useful is the Influenza Virus Resource from the National Center for Biotechnology Information. This website gathers Influenza sequences from the NIAID Influenza Genome Sequencing Projects, and provides access to the data free-of-charge, as well as useful tools for both annotation and analysis of the seqences. Furthermore, you can easily access good information about Influenza viruses, as well as links to other Influenza sequence databases.

Another useful resource is the Fluwiki. This site also brings together many different resources (with links), as well as providing information on Influenza virus and pandemics. The National Instutitue of Allergy and Infectious Diseases (NIAID) also has a good website on Influenza virus, both for lay people and for reserachers.

Sequence Search Methodology

For the study reported here, two different datasets were compiled:  influenza sequences (both nucelotide and amino acid) and sequences recited in patent claims.  The patent sequences were compared to the influenza sequences using BLAST programs. Then, the portions of the patent sequences that had significant homology to influenza sequences was determined.  Criteria for a match varied according to whether the query was with nucleotide or amino acid sequences. 

1. Generation of a searchable influenza genome and protein database

We obtained the most recent influenza genome and protein sequences collections from the FTP site of NCBI's Influenza Virus Resource. There are more than 2,000 complete influenza genome sequences in the database.  Because relatively small changes in the flu genome at either the nucleotide or amino acid level can have large epidemiological effects, all sequences in the database were obtained. The formatdb program from NCBI to convert the data to a searchable BLAST database.

2. Compilation of sequence databases for granted patents and patent applications

Applications

For patent applications, sequences of the bulk sequence applications were obtained from the Publication Site for Issued and Published Sequences (PSIPS) web site. This web site provides sequence listings for U.S. patents and applications that are longer than 300 pages. Sequence listings for the non-bulk sequence listings (fewer than 300 pages in length) are published by the USPTO as an XML document. For each of the listing types (bulk sequence and non-bulk sequence), there was a separate file for nucleotides and amino acids. Data for U.S. applications have only been available since 2001.

The bulk and non-bulk sequence listings were then converted to a common data format (FASTA) and combined to create one database for nucleotide sequences, and one database for amino acid sequences. Additionally, each of these combined databases was converted to a searchable BLAST database for use with CAMBIA's patent sequence search tool.

Granted (Issued) Patents

For granted U.S. Patents, we used a data source that wasn't available for the applications; GenBank at NCBI has a searchable patent database of sequences disclosed in granted patents. To create a database of sequences in granted patents, the U.S. patent sequences were acquired from GenBank, which required removing all sequences that originated from non-U.S. patents.

Sequence listings from the bulk and non-bulk patents were obtained in the manner described above for Applications. The data from all three sources (GenBank, bulk, and non-bulk) were converted to a common format. A filtering step removed duplicate sequences in the data provided by GenBank and by the USPTO (bulk and non-bulk).

The identical process was carried out for nucleotide sequences and amino acid sequences. As with the applications, each of these databases was converted to a searchable BLAST database for use with CAMBIA's patent sequence search tool.

3. Identification of sequences that are recited in the claims of granted patents and patent applications

A key feature of our analysis is determining which sequences are recited in the claims of patents and applications, rather than just disclosed in the specification. To this end, we created four databases that contain only the sequences that are recited in the claims of patents and patent applications. The four databases correspond to nucleotide sequences in applications, amino acid sequences in applications, nucleotide sequences in granted patents, and amino acid sequences in granted patents.

Identification of sequences in claims involves the use of keywords that are used to identify sequence listings (e.g., SEQ ID NO:). Establishing a comprehensive list of keywords is challenging, as many different phrases are used. Applying these phrases resulted in a list of sequence ID numbers that are designated in patent claims. Four new databases were then created that contain only the sequences that are recited in the claims of applications and patents.

4. BLAST search of influenza genome and protein databases using the sequences recited in claims as input

After compiling a collection of sequences that are recited in the claims of patents and applications, we then used those sequences to query the influenza genome database (see step 1) using MEGABLAST to identify nucleotide sequences, and blastp to identify the amino acid sequences that are recited in claims and have significant homology to sequences in the influenza genome.

Nucleotide MEGABLAST parameters:

Positive matches had an E value of 1e^-200 or less, and were at least 150 nucleotides in length.

Amino Acid blastp parameters:

Positive matches had at least 80% identity.

Claims to influenza nucleotide sequences

The patents and patent applications shown in the following tables were identified by a MEGABLAST analysis of the full influenza genome set using sequences recited in granted U.S. patents as the query set. The sequences that had an E value of 1e^-200 or less and were at least 150 nucleotides in length were considered a match. Basic information about each patent document is presented:  publication number, title, assignee (if unknown, the inventor is listed), influenza segment matched, SEQ ID Nos. that matched the influenza segment, and notes summarizing the claimed sequences.

From an examination of the claims in granted patents, several major points emerge:

• One-half of the patents are directed to equine influenza virus (assignee = Heska Corp.);
• None appear to claim the H5 subtype of segment 4, which encodes hemagglutinin (HA) or the N1 subtype of segment 6, which encodes neuraminidase (NA);
• Of the two patents that claim HA sequences, one claims sequence encoding the signal peptide region from HA1 and HA3 (US 6245532), and the other claims an antigenic fragment of HA2 and HA3.

The claims in patent applications present a different picture:

• Claims of three patent applications are directed to reassortment viruses, useful as vaccines against H5N1 (US 20060008473 - MedImmune) or H1N1 (US 20050003349 and US 20070231348 - Wisconsin Alumni Research Foundation);
• Of the remaining patent applications that initially claimed influenza nucleic acids, only about one-half still claim nucleic acid sequences and most of these are directed to sequences encoding internal proteins. 

The tables are current as of 2 April 2008.

Influenza Patent Grants Based on Nucleotide Search

The patents shown in the table below were identified by MEGABLAST analysis of the full influenza genome set from NCBI queried with sequences recited in granted U.S. patents. Criteria for matching sequences were an E value of 1e^-200 or less, and at least 150 nucleotides in length. If the assignee is unknown, the inventor is listed. Related patents and patent applications filed in other countries can be found by clicking on patent number and navigating to the Patent Family and Status tab.  This list is current as of 2 April 2008.

Of the patents listed in the table below, none appear to claim specific nucleic acid sequences encoding a full-length hemagglutinin (HA) or neuraminidase (NA) protein from other than equine influenza virus.  Moreover, fully one-half (6) are directed to equine influenza viruses, while one is directed to swine influenza virus.  Of the remaining patents, a few are directed to sequences useful for vaccine production.  In particular, US 6720409 claims sequences that encode a fragment of HA (H2 or H3 sub-type) that is antigenic.  Although not directly claiming HA and NA sequences, US 7037707 recites a method for producing a reassortant virus in which the origin of the six segments encoding internal proteins is specified and the origin of sequences encoding HA and NA are unspecified.  As such,  the HA and NA could derive any other virus, including from H5N1 virus.  It is not an all-encompassing claim however, because of the limitation on the origin of the other sequences.

Of the 12 granted patents listed below, 9 are assigned to corporations and 3 are assigned to non-profit organizations.

Applications Claiming H5 or N1 or Both

For these applications, PAIR on the USPTO web site was consulted to determine which nucleotide sequences are currently pending.  The applications shown in the table below all claim an H5 and/or an N1 segment.  The MedImmune patent application is directed to a live virus that would raise immunity to H5N1, but which does not cause disease itself.  If these claims are granted (as seems likely), the claims are not very broad; claim 17 requires specific sequences or origins of seven of the genome segments.  The two WARF patent applications are drawn to a similar type of virus to raise immunity to H1N1.

The last column shows the non U.S. jurisdictions where related patent applications are filed. No information is provided here regarding the status of these applications.

Of the 4 applications in the table below, 3 are assigned to non-profit organizations and one is assigned to a corporation.

Influenza Patent Applications to Segments Other Than H5 or N1

The applications listed below recite at least one nucleotide sequence in the claims that has a high degree of similarity to a sequence in the influenza genome set. For each application, we show the number of similar sequences out of the total number of sequences in the published claims.  Although most all of the applications recite multiple sequences in the published claims, there is generally only one or a few claims in prosecution.

The last column shows the non U.S. jurisdictions where related patent applications are filed. No information is provided here regarding the status of these applications.

With respect to policy issues of the H5N1 virus, the sequences in these applications do not relate to the H5 or the N1 subtypes. Moreover, at least half of the applications are either abandoned or prosecuting claims that do not recite sequences.  (Prosecution is the back and forth discussion and argumentation that occurs between a patent office and a patent applicant.)  During the prosecution process claims are often amended or limited, so that until there is allowance of the claims, there is a level of uncertainty as to what will ultimately be granted, or even if a patent grant will occur.

Of the 13 applications listed in the table below, 7 are assigned to corporations and 6 are assigned to non-profit organizations.

All applications in the table had at least one sequence in the original claim set that had a MEGABLAST e value of 1e^-200 or less and were at least 150 nucleotides in length.

Claims to Influenza Protein Sequences

The extent of patent protection for protein and peptide sequences of influenza is equally important to protection for genome nucleotide sequences.  Some types of vaccine production entail use of proteins and peptide, and they are also used in diagnostics.  For these reasons, we searched for and analyzed patents and patent applications claiming proteins and peptides derived from influenza virus.  In particular, patent claims that recite an influenza-derived amino acid (protein, peptide) sequence were found by querying a dataset of influenza protein sequences obtained from NCBI. The minimum length requirement of a match was set to 10 amino acids in order to include claimed peptides.  While the cut-off for inclusion in the following tables is 80% identity, upon inspection of the actual claims, it is clear that a large number of the matches in the 80-95% range are very short peptides and furthermore, were not derived from influenza proteins.  Thus, for the granted patents, length criteria was raised to 95% identity (all exceptions noted in the table); for patent applications, the length criteria remains 80% until a detailed inspection of the claims is undertaken.

From the following data, conclusions include:

• Most of the patent claims are directed to short peptides;
• Most of the patent claims reciting full-length or near full-length proteins are drawn to equine influenza virus proteins;
• Patent claims can be open-ended with respect to sub-types of influenza proteins, such as claiming a reassortant virus in which the sequences of the six internal proteins are specified but the sequences of hemagglutinin (HA) and neuraminidase (NA) are unspecified;

• Peptides in the claims may be used for a variety of purposes, and especially for vaccines;
• Claims of the patent applications are similar to those from patent grants, except that the number of sequences in the claims is higher.

Influenza Patent Grants with Claims to Amino Acid Sequences

Most of the patents in the table below claim peptides, ranging from about 9 amino acids to about 25 amino acids long.  Most of the claims reciting full-length or near full-length proteins are drawn to equine influenza virus proteins.  One patent (US 6337181) is drawn to a method for determining which HA variants are advantageous to the virus by an alignment and comparison of sequences.  Although not directly claiming HA and NA sequences, US 7037707 recites a method for producing a reassortant virus in which the origin of the six segments encoding internal proteins is specified and the origin of sequences encoding HA and NA are unspecified.  As such,  the HA and NA could derive any other virus, including from H5N1 virus.  It is not an all-encompassing claim however, because of the limitation on the origin of the other sequences.

Peptides in the claims are intended for a variety of purposes.  Some recite peptides for use in vaccines comprising peptides (e.g., US 674032), while others recite peptides as part of a fusion protein, such as linked to env protein of HIV (e.g., US 6451322), or peptides that are associated with MHC molecules (US 5976551).

The patents in this table were found by either a blast-p search or a tblastn search. Except where noted, sequences have at least 95% identity to an influenza protein or a translated genome sequence. Most patents were found by both search methods; those found only in a tblastn search are noted.  Claims of patents meeting the criteria were further screened to exclude those that recited an influenza sequence as a minor component of a product and those that recited method claims that were directed to fields other than influenza. 

Of the 20 patent grants shown below, 9 are assigned to corporations and 11 are assigned to non-profit organizations.

Patent Applications with Amino Acid Claims

The applications listed below recite at least one amino acid sequence in the claims that has a high degree of similarity to an influenza peptide sequence. For each application, we show the sequences that are similar to or identical to an influenza sequence.  Although most all of the applications recite multiple sequences in the published claims, there is generally only one or a few claims in prosecution.  The last column shows the non-U.S. jurisdictions where related applications have been filed. No information is provided here regarding the status of these applications.

This set of applications was identified using blastp. The criteria for inclusion in this table is that the patent claim sequences were at least 80% identical to an influenza peptide sequence.

Of the 57 applications in the table below, 40 are assigned to corporations and 17 are assigned to non-profit organizations.