At A Glance
Name: Gilbert Omenn
Position: Professor of human genetics, internal medicine, and public health at University of Michigan.
Director of HUPO Plasma Proteome Project.
Background: Executive vice president for medical affairs and CEO of University of Michigan Health System, 1997-2002.
Member, Fred Hutchinson Cancer Research Center, 1983-97.
Dean, School of Public Health and Community Medicine, University of Washington, 1982-97.
Professor of medicine (medical genetics), University of Washington, 1971-97.
PhD in genetics, University of Washington, 1972.
Research associate with Christian Anfinsen, National Institute of Arthritis and Metabolic Diseases, NIH, 1967-69.
MD, Harvard Medical School, 1965.
AB in biology, Princeton University, 1961.
How did you get involved with proteomics?
I have a long history of research in protein chemistry, going back to my work at the NIH with Dr. Christian Anfinsen in the 1960s. In recent years my interest in biochemical and molecular genetics and the clinical and public health applications has built upon this earlier work. As CEO of the University of Michigan health system, I’ve invested a lot in recruitment and development of programs in the areas related to genetics and functional genomics and proteomics and bioinformatics. So when I completed my administrative job, it was a terrific opportunity to move into a senior faculty role in this domain. In the meantime, with a group of key faculty here and others around the state, we developed the Proteomics Alliance for Cancer (PAC) which was funded by the State of Michigan Life Sciences Corridor. Finally, I became deeply involved with HUPO along with Sam Hanash.
When was the PAC launched, and when did you become involved with HUPO?
[PAC was launched in] August 2002. I was really around from the first meetings of HUPO [in 2001]. This was while I was still full time CEO of the health system.
How do you feel about the progress of HUPO so far, and where does it still need to go?
I think the progress of this organization is far exceeding anyone’s expectations. Usually organizations spend years just getting themselves together, figuring out memberships, titles, people in the council, holding some scientific meetings — all of which is good. But this organization decided very early that it was important to mount major scientific initiatives that would advance the field and attract governmental and corporate investment.
What are the next steps in pushing HUPO forward?
The scenarios for the organization have a lot to do with the scenarios for the advancement of the field. If governments and corporations mount substantial initiatives, HUPO would play an influential role [even] if it were only a role of annual congresses, regional meetings, publications, and facilitation. The bolder goal for HUPO would be to actually organize large-scale initiatives in place of, or before, governments might do it themselves. That is what is happening with the liver project and is still emerging with the brain project. With regards to the plasma project, HUPO has very nice pilot phase funding from industry and the NIH. The challenge is how to move into a large-scale initiative that would look at the discovery of protein biomarkers and monitoring protein biomarkers in health and disease. That’s a very big, very broad agenda which has the same kind of problem that the human genome institute has at the NIH, which is that as it succeeded, it moved into the turf of many disease-specific institutes. So a lot depends on how the NIH proceeds. I’m very proud that they have come together across six institutes of the NIH, with some trans-NIH funding for the HPPP.
It looks like the pilot phase of the HPPP will have substantial value for future studies of the proteins in the blood and studies focused on tissue organ proteomes, for each of which it would be desirable to have parallel analysis in the blood to see what biomarkers emerge.
You set a data submission deadline of Dec. 31, 2003 for the pilot project. Is all the data in now?
It’s not quite final. We’ve really pushed hard on the timelines for this, because people have the results, and because some of our major contributors of data generously offered that they would hold publications so that the whole project could be published together. Obviously that puts an obligation on everyone else to get their work done more expeditiously. What there will be is multiple individual lab papers, and then quite a few cross-cutting analyses by comparing different kinds of specimens, preparation, depletion methods; results within related technologies across laboratories, across technologies, and how the data repository is established, how it functions, what different search engines do — we should have a tremendous amount of information.
After the Montreal [HUPO Congress] meeting, we set the end of December for full submissions of data. We have quite a few of those now, and we’re poring through what’s come in. We’re quite pleased with what’s been submitted. In many cases, labs have additional specimens [that they’re looking at now]. We have the same situation in our own lab: How deeply do you mine the proteome of one specimen, as opposed to looking substantially but not as deeply across many specimens. We distributed up to 17 reference specimens, and labs could request anywhere from one to 17. At this point, no lab has done all 17, [but] one lab has done 12, and they said they would be willing to do the other five — they just couldn’t get them all done before Dec. 31.
What kind of variation in techniques have you been seeing among the different participating labs?
There are several labs that used the direct MS or SELDI method, where we’ve got to come to grips with the lack of protein identification. We hope that one or more of those labs would be able to couple their analysis with further MS/MS to identify proteins.
In Montreal we also agreed that we had raised sufficient funds [of $200,000] to offer small grants for special projects on top of this tremendous amount of baseline work, which is volunteer. [The projects] generally are looking across technologies: We’re examining in depth different depletion methods and advancing complex new ways of analyzing the data and gaining confidence about protein identifications via such projects. We expect those projects will be completed by mid-May.
We are planning a jamboree-style workshop in June. The idea would be to be very well-planned in advance to make maximum use of time. We would have about 20 work stations and have technology leaders and bioinformatics experts examining what is in the database, [and] asking fresh questions — cross-cutting questions that can be answered with intensive collaborative work face-to-face. At the same time [we’ll] have stations for manuscript production so we can generate the manuscripts, put them through initial review, and not have a dribbling in of manuscripts delay the publication.
What needs to happen between now and the HPPP jamboree in June?
First, [we need to] complete the importation of the datasets that are being sent to us; second, expand the analytical work and submissions over the next few months beyond what was actually completed in time for Dec. 31. Third is the award of these small grants, and fourth, the accomplishment of the special projects analyses.
There’s another feature that links the pilot phase to the future. While we’re immersed in what I was just describing, we need to work with the NIH and with governments in other countries and with the several companies that have already been sponsors to define the most appropriate path for developing pilot proteome work in health and disease going forward. Our long-term goals are to comprehensively analyze the protein constituents of plasma serum; to characterize changes that occur in people as a function of gender, age, diet, and various other factors; and to identify and apply proteins as biomarkers of disease processes with relationships to many different organs. One aspect of that is assuring that there’s a plasma component to the liver and brain proteome projects, and to many other organ-based projects — such as salivary gland, and cardiac — which have been recently proposed to us.
Who is organizing all the pilot project data?
The primary informatics components are at the University of Michigan, led by David States, and the EBI, which has the lead for all of HUPO on bioinformatics and databases. So that’s the operating combination for handling all the data, feeding back queries to the investigators, and advancing the production of the database, all of which will be made open-access. Then, we have proposals from some other investigators as part of special projects and part individual participating labs for extensive analysis across datasets.
What is your own lab working on?
We are very keenly interested in development of biomarkers in blood for much earlier diagnosis of common cancers. The five-year survival for lung cancer after diagnosis is only 15 percent — really not better than it was when I graduated from medical school, despite some good efforts. We have some promising markers from work done with Sam Hanash here at Michigan. We think we can generate several more that would be able to detect lung cancers early. We have specimens from people before they develop lung cancer and people who did not develop it, with a lot of follow-up. There’s also work at Michigan involving Craig Logsdon and Sam Hanash on pancreatic cancer. And further on, colon, and other sites of cancer. The general approach is to identify markers for early detection of many types of cancer.