By John S. MacNeil
If you talk to biologists about the current model for generating robust and high-throughput genome sequence and DNA microarray data, most would agree that core facilities are a fairly reliable source. As sequencing and array analysis technologies have increasingly become commodities, service laboratories that provide these data have in recent years begun to take advantage of economies of scale by joining forces across regions or forming mega-facilities. Nowadays most genome sequence data come out of the “Big Three” genome centers: Baylor, the Broad Institute (formerly Whitehead), and Washington University. And organizations such as AMDeC in New York have banded together their microarray core facilities to share expertise and increase their bargaining power with vendors.
But what about proteomics? Are regional core labs the solution to providing high-quality protein-analysis data at a reasonable cost? While some researchers are giving it a shot, a look at the state of the field shows that the answer is a fairly straightforward “not quite yet.”
Compared to sequencing and microarray analysis, the technology for proteomics experiments is at a much earlier stage of development. Given the number of competing platforms and approaches — even for such types of analysis as identifying proteins by mass spectrometry — there just isn’t a consensus for what specific platforms a regional proteomics core facility should include.
“I would say that proteomics is technologically similar to where genomics was in the early ’90s,” says Phil Andrews, a proteomics researcher at the University of Michigan. “But that’s a simplistic way of looking at it because proteins represent a much greater analytical challenge than DNA sequencing. For one thing, we’re trying to measure a variety of different properties of proteins … so it can be quite messy in terms of the variety of technologies that one is using.”
The scientists who lead the most advanced proteomics labs in the US — Dick Smith at Pacific Northwest National Lab, Andrews at the University of Michigan, and Ruedi Aebersold at the Institute for Systems Biology, among others — are all striving to balance technology development with their ability to serve their research constituencies. But that’s not to say they’re all taking the same approach. How their models differ for shepherding proteomics through its adolescence will have a profound impact on what the field looks like when it grows up.
The Old Model at Work
To be sure, there are certain types of proteomics analysis that are pretty well hammered out. It doesn’t require an FT/MS system, for example, to simply identify most proteins in a relatively well-characterized system such as E. coli. A liquid chromatography system hooked up to an ion-trap spectrometer may be sufficient to handle many run-of-the-mill experiments. But unlike DNA sequencing — or even DNA microarray analysis, for that matter — proteomics experiments can vary widely in their complexity, as well as the ability of the technology to reduce that complexity to a simple solution. Researchers using high-end technology see the potential for vast improvements, and are hungrily seeking the next advance.
Simultaneously, researchers from across biology are putting pressure on the technology leaders to share their expertise. An informal survey at the recent annual meeting of the Association of Biomedical Research Facilities showed that scientists who run core facilities for protein analysis are experiencing ever growing demand for their services. Meanwhile, disciples of mass spec such as John Yates at the Scripps Research Institute are overwhelmed with requests from researchers who want to visit his lab and learn protein mass spectrometry.
Satisfying these dual desires of production and new technology can take different forms. In some cases, prominent proteomics scientists and their funders have attempted to create regional facilities, where pooling resources allows one laboratory to purchase cutting-edge technology or instrumentation in return for an obligation to provide research services for the biologists within a specific geographic area — most often one state.
Andrews at the University of Michigan serves as the leader of one such initiative. In 2001, flush with the proceeds of a successful tobacco lawsuit, the state of Michigan established a fund to support biotechnology research across the state. As part of this endeavor, the state created the Michigan Proteome Consortium, with Andrews as its director.
The purpose of the consortium, Andrews says, is primarily to provide proteomics research services to academic and corporate labs located in the state. The facilities are anything but centralized, however: in concert with colleagues at Wayne State University, Michigan State, and the Van Andel Research Institute, Andrews’ consortium offers services ranging from proteome mapping, protein microarray analysis, yeast two-hybrid analysis, and protein informatics.
Andrews says he’s also been lucky enough to receive funds from the NIH’s National Center for Research Resources to help support the consortium’s efforts to develop new proteomics technology. Funding from NCRR allows him to spend $1.7 million annually on new technology R&D while still maintaining a primary focus on the consortium’s service component.
Another attempt to create a regional center for proteomics technology and research services is taking shape in Indiana. With $2 million from the 21st Century Research and Technology Fund, an entity established by the state of Indiana in 2001, and $3.2 million in funding from Eli Lilly, the Indiana Centers for Applied Protein Sciences have a mission to provide technology validation, protein analysis services, instrumentation, and technical support for both academic and industry investigators in the state. The current head of proteomics programs at Lilly, James Ludwig, will lead the center as founding CEO.
Other proteomics big shots are appealing to a national or international audience. Rather than focus on serving researchers in a particular geographical area, these scientists say a more useful approach would be to create centers with the resources to design systems capable of handling the most rigorous proteomics experiments, in terms of both throughput and analytical precision.
Aebersold at the Institute for Systems Biology in Seattle has for several years advocated funding to support “national centers for proteomics” modeled after high-energy particle physics labs with cyclotrons or telescopes at astronomical observatories. The idea, he says, is “to address the need that clearly is there in the biological community to go and have access to collecting high-quality data on quantitative proteomic experiments.” Where the facilities are located, he adds, is secondary to the lab’s scientific capabilities and universal access.
Although Aebersold’s vision has yet to result in funding from the NSF or NIH, there is evidence that the idea is gaining traction. Aebersold cites Smith’s lab at PNNL as one laboratory with the resources and expertise to build proteomics instrumentation on “the bleeding edge,” as they say, of technology development. The new laboratories under construction at the headquarters of the Howard Hughes Medical Institute in northern Virginia might also serve as a locus for high-end proteomics research.
Speaking with Smith at PNNL provides some insight into how such a technology center might operate in practice. In his protein mass spectrometry lab, there are two instrument platforms: an assembly line for churning out protein identifications from highly complex samples using his production-scale Fourier-transform mass spectrometry system, and an R&D FT/MS system for developing new experimental protocols and technology. Every few months, Smith decides that elements of his R&D platform have proven worthy for transfer over to his production proteomics platform.
“Things change rapidly,” he says. A periodic upgrade model “is important at present because the technology is not at a mature level.”
Perhaps that may not always be the case. Given the amount of resources currently devoted toward expanding the reach of proteomics technology and ironing out the kinks in its routine application, there may soon be a day when acquiring high-quality protein analysis data may be as easy as ordering up sequencing data. Until then, program managers at funding agencies will remain on the sidelines. “It’s early days,” says Doug Sheeley, director of the Division of Biomedical Technology at NCRR. “You can’t provide routine access to technology that’s not fully developed.”