Summary of ISGO Task Force Meeting
Berlin, 10-October-2002
(Notes by Tom Terwilliger; please report any errors to "terwilliger@lanl.gov")
Dr. Udo Heinemann (Berlin, Germany) opened the meeting with a brief summary of the history of structural genomics, the role of international cooperation and coordination, and the role of the International Structural Genomics Organization.

Dr. Norihisa Hara (Ministry of Education, Culture, Sports, Science and Technology, Japan) gave a presentation describing the importance of structural genomics, the fact that structural genomics is basic research, the role of international cooperation, and the importance of data release for promoting scientific research. He mentioned the importance of having a possibility for a 6-month hold on coordinate release, allowing time for patenting.

Dr. Thomas Terwilliger (USA) talked on behalf of Randy Read (UK) about the work of the committee on "Numerical criteria for evaluating and assuring data quality". The fundamental conclusion of this committee was that there should be no significant lowering of quality standards for structural genomics compared to any other structural biology work. Further the committee felt strongly that it was essential to deposit fundamental data (i.e., structure factors for X-ray; distance constraints for NMR) along with coordinates. Discussion included the importance of defining "resolution" in addition to the other criteria already discussed by the committee.

Dr. Stephen Bryant (USA) talked about the work of the task force on "Tracking and Registration of Targest" and the resulting target tracking database. This is an important role of ISGO, helping foster international coordination by letting researchers in structural genomics know what targets each group is working on and what progress has been made. The committee set up a list of status indicators (such as "targeted", "purified", "in PDB" that give a general idea of the status of each target. This has been fully implemented through the PDB and is in constant use. Audience discussion of the database included comments that a fuller biological description of the targets would be useful (i.e., what domain/quaternary structure is targeted, not simply an ORF name).

Dr. Helen Berman (USA) described the work of the task force on Deposition, Archiving, and Curation of the Primary Informationｲ. This is a key task force that has spent a great deal of effort defining just what information needs to be collected, how to go about collecting it, and implementing it. A new set of data dictionaries (what is to be collected; definitions of all these items) have been developed. Dictionaries for X-ray and NMR data are now well-developed, the dictionary on protein production is drafted and is being tested. Dr. Berman described an off-line version of the PDB deposition data-checking software (ADIT) that anyone can use to pre-check their data format/content before deposition. Dr. Berman listed a number of web sites providing useful information on deposition for structural genomics (see presentation). In the discussion, Dr. Berman emphasized that the targeting database records status only; the PDB collects final results. Another possibility is that the PDB might archive data that has value but has been abandoned (i.e., x-ray data that did not lead to a structure). This could be done as a matching-service listing datasets that exist and where to get them. She felt that at this point the PDB wasnｹt currently ready to take on the mission of archiving all the data for failed experiments, but was ready to hear the voice of the community on this. Dr. John Markley pointed out that many soluble proteins have unstructured regions or are unstructured, and that the BMRB is collecting this information. Additionally Dr. Berman mentioned the PDB default for pre-release of sequence information for proteins on hold is now "release" and is hoping that this will encourage the release of this information. Further discussion involved the idea of links in the deposited data to the biology of the proteins.

Dr. Guy Dodson (York, UK) described the work of the task force on "Publication and Recording of Methods". This task force after much debate concluded that publication of results from structural genomics was indeed desirable. The question is how to report (describe) structures when many are being done, including biological relevance. The committee felt that short descriptive papers would be very useful. Several journals have been quite responsive in this area, including in particular Acta Crystallographica, and the Journal of Structural and Functional Genomics. Dr. Dodson concluded that the committee has largely finished its mandate, that now it is only necessary to check up on the progress of short publications. Gaetano Montelione pointed out that Proteins also is encouraging "structural notes" suitable for structural genomics. He pointed out that when contributors put in a lot of information, the papers become more like a regular paper and require a lot of review work. Their solution is to limit these short notes to 3 pages. There was a lot of discussion about the role of in-depth papers for structural genomics and the fact that most papers in the field are in this category. Mitchell Guss pointed out that Acta Crystallographica is working to provide technical validation data to reviewers automatically, and that this could possibly be extended to biological information.

Dr. John Norvell (USA) discussed the work of the task force on Intellectual Property, including the Hinxton, UK and Airlie House, USA meetings in which the goals of free exchange of data and timely deposition were discussed, and in which policies for data release were developed. He emphasized that the support of structural genomics varies from place to place so that policies needed to be designed to accommodate differences. The Airlie House meeting resulted in an agreement for deposition of coordinates immediately on completion of structure determination, with public release in a short time (but always less than 6 months). Dr. Norvell pointed out that the Airlie agreement emphasized the desire of the organization to have stringent requirements for utility in patenting.

Dr. Barbara Skene (Wellcome Trust, UK) discussed a proposal for an addition to the Airlie agreement. The Wellcome Trust was worried about the 6-month delay which was allowed in occasional circumstances in the Airlie agreement and which was designed to allow patenting. Some people were worried that any patents that were issued might impede academic research. Dr. Skene proposed that an addition to the Airlie agreement that might be useful would be one that specifies that patents on coordinates of structural genomics-based structures should grant some type of exemption for research/non-profit research uses.

Dr. Lawrence Sung (University of Maryland, USA) discussed the impact of patent rights on the dissemination and use of structural genomic data. He discussed patent protection of genomic inventions, research exemption, and the ISGO draft data release policy. Note: the US has a different policy on "non-commercial uses" than other countries. He discussed US policy principally. Conditions for patentability include subject matter (rather broad, however note that information itself is not patentable). Other conditions are utility, novelty and non-obviousness. The utility must be specific., e.g., a "dsDNA sequence as a probe" is not sufficient. For sequences, the utility had to be specific and credible; ESTs without known relevance are now being held up in the US patent office. US patent office currently has a guideline that says they will not reject homology-based arguments for utility (i.e., if sequence A has 90% identity with B, and B has a clear utility, then A has utility). In Dr. Sungｹs opinion, it is unclear if this will stand up. Disclosure requirements in the US include definiteness, written description (description should convey that applicant is in possession of the invention), enablement (without undue experimentation), and "best mode". The description is of a chemical (DNA sequence is part of the chemical description). The US has no research exemption (i.e. exemption for non-commercial uses). There are mechanisms to guard against practical effects of non-commercial uses (why would a company sue an academic for a non-commercial use; also federal and state agencies may not be sued for patent infringment; also US government grantees have to provide the US government with rights). The only ones left out are private institutions at this point. Dr. Sung discussed the proposed ISGO data release policy and providing something like a research exemption. One way is a license provided by the patent holder. Another is a covenant not to sue provided by the patent holder. Discussion included comments on structural patents. Dr. Sung pointed out that a coordinate set by itself is not patentable, but if there was a process set out for using the coordinates it could be. A question was what about patenting an active site? The answer is the utility must be described in detail; i.e., what is the site good for?

A panel discussion continued on the patenting issue. John Stewart (Wellcome Trust) commented that the question of patentability of protein structures themselves appears to be in dispute. It seems that protein structures themselves cannot be patented but uses of them might be. Dr. Sung responded that "you can always apply for a patent", but what counts is what would happen after issuing. Structural coordinate data would not be patentable. Further a methodology in standard use or obvious applied to coordinate data would also not be patentable. Only a new use would be, and the coordinates themselves would not be patentable. Sara Vinarov pointed out that there were various types of claims, "using 3d coordinates to identify agonists", "using 3d structures to enhance bioactivity by modifying specific residues", "using 3d structures to enhance stability". She pointed out that the research exemption is highly controversial, but that there is effectively no research exemption in the US. Dr. Hurtenbach said that all data obtained by public funding from the German Ministry of Education and Research should be exploited (patented if appropriate, published if basic research). She pointed out for example that data from the Protein Structure Factory would be published with very few exceptions that might be patented. She suggested a continuation of the grace period of 6 months. Dr. Enfedaque (EU) pointed out that there is an EU commission (led by Frederic Marcus, EU; frederick.marcus@cec.eu.int) studying IP rights in the biotechnology area. She described some of the policies under study. Dr. Hara (Japan) pointed that IP issues are not just for structural results, but also for sequence. Tom Terwilliger (USA) proposed a simple statement for support by the ISGO membership, that, "ISGO strongly supports the position that a high level of utility should be required for patenting of protein structures," noting that ISGO continues to expect that most structures will be
immediately deposited and released, as agreed at Airlie House. This statement will be submitted to the ISGO membership for consideration.