How will the soon-to-be incorporated HUPO champion proteomics in the greater scientific community? A crowd of 300 gathered to find out in San Diego Jan. 9-11 at a follow-on to the slightly better-attended inaugural meeting that was held last year in Tysons Corner, Va.
But except for an opening talk by HUPO President Sam Hanash, most of the meeting’s discussions concerned technologies and other advances that might prove applicable to a larger proteomics project, in whatever form it eventually takes. In addition, prominent researchers debated various technical issues confronting the field in several panel discussions. Following are some highlights, tidbits, and asides from the conference:
Liquid ICAT or Solid ICAT?
Ruedi Aebersold of the Institute for Systems Biology spoke about his group’s recent efforts to devise a solid-phase assay using his widely-touted ICAT reagents for measuring changes in protein expression between two samples. Attaching the ICAT reagent and reactive moiety to a glass bead simplifies the purification procedures, he said, and results in higher yields.
As an example, Aebersold cited an experiment that compared yields from the conventional liquid-phase ICAT reagent with that of the ICAT reagent attached to a glass bead. Using 2.5 micrograms of protein taken from a yeast cell lysate, Aebersold said the solid phase assay found 57 proteins with discernible variations in protein expression, whereas the conventional ICAT reagent found only 18. Thirteen proteins were detected by both methods.
In another advance, Aebersold described his group’s progress in developing mass spectrometry techniques for increasing the throughput of experiments using the ICAT reagents. The strategy, he said, is to “reduce the noise from proteins that do not change significantly in expression” by selecting for amino acid sequencing only those proteins with differences in expression that pass a certain threshold.
To do this, Aebersold and Chris Lock, his collaborator at MDS Sciex, have designed control software for Sciex’ QSTAR instrument that allows the peptide mass fingerprint spectrum to be used as the basis for selecting which tagged proteins to sequence in MS/MS mode.
In a panel discussion aimed at comparing the relationship between mRNA and protein expression data, the conversation soon led to a debate on the need for ways to share protein expression data between researchers. Walter Blackstock, Cellzome’s vice president for technology, expressed the desire to analyze other researchers’ expression data for evidence of proteins that cannot be immediately identified through database searching algorithms. John Yates of the Scripps Research Institute countered in jest, “Can I deposit all of my data on your computer?”
Jokes aside, Yates suggested that researchers develop a file transfer mechanism, akin to Napster’s now defunct MP3 file-sharing software, that would allow colleagues to access the raw data from mass spectrometry-based protein identification experiments. A similar “peer-to-peer” file transfer protocol, with safeguards for protecting the integrity of the underlying data, could provide a solution for researchers wishing to contribute their expertise to correctly identifying proteins, he said.
In a broader sense, such a scheme for file sharing would open the door for issues related to ownership of the raw data, added Andrew Link, a yeast proteome researcher at Vanderbilt University. “Who gets the credit?” he asked. “Those who analyze the data or those who generated it initially?” The answer boils down to the conflict between scientists’ ethical commitment to make data public as soon as possible, and the need for the same scientists to achieve tenure, commented John Wooley, associate vice chancellor for research at the University of California, San Diego.
As an example, Wooley described the case of a researcher in the 1950s who generated data in experiments involving Parkinson’s disease, but died before he could analyze the results. Twenty years later, his colleagues ultimately recreated the experiments, which led to advances in understanding the disease. “What if that data had been available before?” Wooley asked rhetorically.
We All Scream for Flexgene
To considerable interest from conference attendees, Josh LaBaer, director of the Harvard Institute of Proteomics, described his progress leading the development of the FLEXGene repository, a planned expression-ready collection of full-length cDNAs representing every coding region in the human genome. LaBaer said his group can currently produce about 500 clones a week, and plans to release the first version of a repository covering every full-length coding region in the S. cerevisiae genome within the next several months.
Completing the repository for the estimated 35,000 open reading frames in the human genome within a reasonable period of time will require, among other things, significant advances in throughput, LaBaer said. “How can you do this in high throughput?” he asked. “The answer is, we’re getting there.” LaBaer’s efforts to scale up the production of cDNA clones and verify their sequence will be published in a forthcoming issue of the Proceedings of the National Academy of Sciences.
Once LaBaer completes his collection of human cDNAs, he plans to license rights to express the clones to any company “as long as they show they can do it carefully,” he said, mentioning Resgen and The Image Consortium as possible partners.
LaBaer also pitched the FLEXGene project as a potential candidate for sponsorship by HUPO, because it addresses many of the criteria the organization has stated warrant its support. The FLEXGene repository will be validated, comprehensive, and involve the input of both public and private research groups, he said.
Although funding for the initiative currently comes predominantly from public funding agencies, LaBaer has enlisted several undisclosed companies to serve as advisors and potential financial backers. LaBaer has said he would need about $100 million to build a comprehensive collection of 100,000 clones.