NEW YORK, Jan 31 - Can sequencing do for the proteome what it did for the genome?
On Wednesday, a number of world-renowned researchers in the field of proteomics issued a resounding " no."
Instead of devoting their efforts to decoding the human proteome, proteomics researchers should focus on developing a larger picture of protein structure, function, and pathways within cells and organisms, panelists said at a New York Academy of Sciences briefing entitled “The Promise of Proteomics."
“When a company has phenomenal success with strategy A, you want to do strategy A on the next subject,” said John Richards, a professor of organic and biochemistry at California Institute of Technology, referring to current corporate attempts to map the proteome.
“This doesn’t work,” he said.
The reason that the sequencing-database model won’t work for proteomics, said Richards, is because proteins are inherently much more complex than genes.
For example, he noted, genes encode multiple proteins through the procedure of alternative splicing, generating multiple different mRNAs for one DNA sequence. “But for all of Celera’s computations, they couldn’t tell you how alternative splicing is going to happen,” he said, adding that genes can also modify protein structures once they are made.
Richards illustrated his point by showing a slide of a Norman Rockwell painting in which a mass of strangers huddled on a train platform waiting for the commuter train, followed by another slide in which a small group of people interacted with one another. The former painting, he said, resembles an assemblage of proteins isolated on 2-d gels, where the relationship between the proteins remains unknown, but the latter painting, in which relationships are evident, is what scientists involved in proteomics should be aiming to do.
Richards said he had told his colleagues at Celera, whom he had helped out at its founding, that they should focus on relationships and function, but he said they didn't listen to him and still remain committed to their strategy of developing a human proteome database.
The Applera foundation, a non-profit set up by Celera’s parent company, sponsored the proteomics panel.
Taking a stab at another popular proteomics approach, Brian Chait, head of the laboratory for mass spectrometry at Rockefeller University, said that high-throughput x-ray crystallography also misses the mark since protein structures don't offer an understanding of cell function.
“We would like a molecular microscope where you can peer into a cell and look at how things work,” Chait said.
MALDI/TOF mass spectrometry, Chait asserted, will allow function determination. The process of taking proteins and digesting them, then measuring the masses of the parts by beaming a laser out on the protein matrix and measuring the time of flight to a charged ion reflector, can help scientists see how proteins are built.
“Then [scientists can] reconstruct a 3-D image of molecular structure for the machines and look at how the molecular machine works,” he said.
Other panelists noted that mass spectrometry, which is currently being developed for high-throughput proteomics by Marvin Vestal and Steve Martin at Applied Biosystems, still has its limits, and X-ray crystallography does have value.
“Proteins are too complex to have one single technology work for all proteins,” asserted Denis Hochstrasser, professor of medical biochemistry and pharmacology at the University of Geneva in Switzerland and a founder of Geneva Proteomics.
Hochstrasser said a few scientists are currently attempting to organize a Human Proteome Organization, HUPO, that could coordinate proteomics research in a similar way to the Human Genome Organization’s coordination of genomics work.
But, he commented, there is less of a need for such a coordinating body in proteomics than in genomics because there is less danger of waste from overlapping efforts; unlike a single string of genome sequence, three-dimensional and highly plastic proteins lend themselves to multiple efforts, he argued.
Joshua Lederberg, president emeritus of Rockefeller University, added in a final note that even if human proteins are completely pinpointed and mapped, these proteins make up only part of the proteome living within human beings.
“We carry with us externally and internally, a vast array of other genomes from pathogens, viruses, et cetera," he said. " Any mapping of the human protoeme has to take these into account as well.”