A database of genes that make normal cells go awry and turn cancerous was formally unveiled this week by the National Center for Biotechnology Information (NCBI). SAGEmap, as it's called, is the first of several gene expression databases in the works.
Watching genes blink on and off is a red-hot research area for studying everything from strawberry ripening to how viruses cause disease (Science, 15 October, p. 444). But researchers are still sorting out how to share their data. Now one effort, the Cancer Genome Anatomy Project, has launched the "first truly public gene expression database," where researchers can both contribute and download data, says Duke pathologist Greg Riggins, whose team describes the project in the 1 November issue of Cancer Research. Using a DNA sequencing technology called SAGE, team members have found, for example, 471 genes that are expressed differently in brain tumors and normal brain cells. You can also type in a gene name to get a "digital northern": data for how that gene is expressed in the cells SAGEmap has studied.
Scientists are struggling to assemble gene expression databases for all techniques, particularly DNA chips. One obstacle is that unlike SAGE, these methods tell you how much a gene is expressed relative to other genes, so it will be hard to compare results across experiments. To find out how experts will solve this problem, keep an eye on NCBI's page for its Gene Expression Omnibus, to open next spring; and on a plan being crafted by the European Bioinformatics Institute.