The San Diego Supercomputer Center, a national laboratory for computational science and signaling, will manage data and develop new bioinformatics software for the Alliance for Cellular Signaling, a consortium of more than 20 academic institutions to compile data on cellular communication in mouse heart muscle and immune system cells.
“We will have a bevy of experimental data coming our way,” said Shankar Subramaniam, an SDSC fellow and a professor of bioengineering at the University of California, San Diego, who will oversee the center’s contribution to the alliance. Proteomic, mass spectrometer, gene array, DNA microarray, and sequence base data are some of the information types the team expects to handle and make available for searching and analysis on the Web.
AFCS, a group of some 50 scientists, will focus on cellular communication for mouse heart muscle cells and B cells from the mouse immune system. These cells communicate in and between themselves by molecular processes almost identical to those in humans.
The alliance will enable SDSC to take the next step to understanding how genes work together to produce certain effects. “We will be finding pathways and networks of pathways,” Subramaniam said.
Mouse Genome
This project along with the complete mouse genome should aid drug discovery research as scientists will be able to view the precise genetic makeup of cellular communication, according to AFCS.
Of the two efforts to sequence the mouse genome, Celera Genomics expects to be finished by year’s end but its data will remain proprietary. The other project, a public and private consortium involving the US National Institutes of Health, Affymetrix, SmithKline Beecham, and Merck, plans to be done in February and will make its data publicly available.
SDSC will also develop bioinformatics software for data analysis. For the signaling alliance SDSC expects to integrate publicly available analysis tools into new software, which may be modeled on the existing SDSC Biology WorkBench software, a Web-based protein and nucleic acid database search tool.
“But some analyses are going to be specific enough that we will have to completely develop programs on our own,” said Brian Saunders, a bioinformatics program analyst in Subramaniam’s group. The system will run on a Unix platform on Sun Microsystems machines.
“Operating the database may take up two terabytes of space, and possibly more,” Subramaniam said.
Subramaniam and his team will create a system that only requires a Web browser like Netscape or Microsoft’s Internet Explorer to enter research data in the databank. Since many different kinds of research will be done, one of the main challenges is to find a system that can handle and parcel out many different kinds of data.
The tasks of receiving, storing, and processing data will be divided among two groups, the Data Coordination Core laboratory and the Bioinformatics Core laboratory at SDSC, which is located at UC San Diego. The supercomputing center hosts one of the world’s largest data storage systems, said Subramaniam.
The alliance will rely on SDSC for the construction and maintenance of its Web presence, and to organize the many types of data that will come from the alliance’s partner institutions.
The AFCS project is an ambitious undertaking because thousands of signaling molecules have been identified in cells, and assembling a catalog of all of the possible and actual interactions among them will not be a simple task.
Instead of maintaining intellectual property rights to research results and publishing in peer-reviewed journals, AFCS will publish data analyses on its website for the research community to examine and duplicate.
Alfred Gilman, a pharmacologist at the University of Texas Southwestern Medical Center at Dallas and leader of the AFCS effort, said that the researchers hope to be able to create a piece of a virtual cell. This could help drug developers get closer to their goal of being able to test new compounds in silico.
The alliance is funded in part by the NIH’s National Institute of General Medical Sciences, which is supplying $5 million in direct costs and another million for facilities and administration. “We’re trying to put data in the public domain, and looking for large collaborative efforts,” said Rochelle Long, chief of the Pharmacological and Physiological Sciences branch of NIGMS.
In addition to the NIGMS money, several pharmaceutical companies and non-profit organizations including Eli Lilly, Johnson and Johnson, the Merck Genome Research Institute, Novartis Pharmaceuticals, Chiron Therapeutics, Aventis, and the Agouron Institute are providing funds.
Myriad Genetics of Salt Lake City and Isis Pharmaceuticals of Carlsbad, Calif., will participate by providing antisense reagents and a two-hybrid analysis technology, used to track interactions between proteins inside cells.
The California Institute of Technology, the San Francisco Veterans Administration Medical Center, and Stanford University are also part of the alliance.
—Martha Heil