On page 349 of this issue, Daniel Toma and colleagues describe an elegant application of microarray results that moves from (polygenic) phenotype to candidate genes, and back to (monogenic) phenotype. The authors compared the gene-expression patterns of two populations of Drosophila melanogaster artificially selected for differential responses to gravity—one population prefers to crawl toward gravity, and the other, away from it. They confirmed their results by testing flies, each of which had a mutation in a candidate gene. (Forty years ago, the same strains were used to first demonstrate that a complex behavioral trait can have a genetic component; see the News & Views article by Steven de Belle on page 329 for further comment.)

One important reason Toma et al. were able to bring their results full circle, from a polygenic phenotype to the confirmation of single contributing genes, is that the Drosophila research community has a strong infrastructure that includes a clearinghouse (called FlyBase) for detailed information on genes and mutants. The authors used this database to track down the genes whose expression was altered in their two populations, and to identify fly stocks carrying mutations in those genes. FlyBase includes, but is certainly not limited to, information on genes, mutations, expression patterns and anatomy. The database also has a cooperative relationship with the Bloomington Drosophila Stock Center—which houses close to 10,000 fly stocks available to researchers for a nominal fee—and maintains information about many other mutant strains available from individual laboratories. The successes of FlyBase and the stock center can be attributed to a history of cooperation among Drosophila researchers, who have been freely sharing stocks for almost a century. The stock center originated in the small room at Columbia University where Drosophila genetics was born through the pioneering work of T.H. Morgan and colleagues. Their policy was simple: any researcher requesting the stocks received them. The stock collection slowly grew and was moved first to the California Institute of Technology and then to the University of Indiana, where it has expanded substantially.

There is no reason to think that researchers working with flies are immersed in some collective love-in. On the contrary, the fly community is as aggressive and competitive as any other, and fly researchers are in no danger of being considered wild altruists. But, although there are certainly some people who do not freely share their materials, there seems to be a general recognition among the community that individuals have much to gain by supporting the collective infrastructure. They recognize the advantage of working together to gain financial support so that individual labs can compete not just with each other, but also with those studying other model organisms.

The success of FlyBase and the Bloomington Drosophila Stock Center is unparalleled among research community databases. The rich history of cooperation in the fly research community has certainly added to this success. Many other communities have similar resources that are based largely on the FlyBase model: for example, Wormbase and the Caenorhabditis Genetics Center (CGC) for the nematode research community, the Zebrafish Information Network (ZFIN) and Zebrafish International Resource Center, and The Arabidopsis Information Resource (TAIR) with the Arabidopsis Biological Resource Center.

In contrast, many other research communities have not created an effective infrastructure that will ensure broad success across multiple disciplines. This probably has as much to do with politics and egos as it does with money and community organization. The stumbling blocks encountered are often viewed as insurmountable, especially in the mouse and human research communities, where freely sharing stocks and samples after publication is sometimes seen as an option rather than an obligation, despite the fact that researchers are required to do so to publish their research in a reputable journal. Stumbling blocks aside, it is clear that to truly capitalize on the information gained during this era of genomic catalog biology, individual communities will have to create dynamic infrastructures and eventually link them together. Building an infrastructure requires public funding for large-scale databases with curators, stock centers for animals, cell lines and DNA libraries, grant incentives to address the technical limitations of the individual communities and a cooperative effort by all who stand to benefit. This is not to say that everyone should go out and create their own database of genes or mutant strains or expression patterns for the community to use. This is not the point and is, in fact, a big problem in some communities. It reflects general disorganization and a lack of cooperation. The fly community has set a fine example of how small and large labs alike have benefited from the common sense of a few people in a small room on the Upper West Side of Manhattan. Other communities have followed this lead, but there is still an immediate need for similar infrastructures to be built and properly curated in all research communities.

Finally, it is convenient to speak of research communities based around the model organism, disease or biological process being studied, but the artificiality of such divisions is becoming increasingly noticeable. A gene implicated in one process in one system may well have a different role in another system, as illustrated by the pleiotropic nature of the genes identified in the study by Toma et al. As individual community resources and databases become larger and more specialized, it will become more difficult to place a genetic interaction in context with all that is known of the genes' function, expression pattern and variations. The Gene Ontology consortium (Nature Genet. 25, 25–29, 2000), as well as large databases such as OMIM, has taken steps to integrate information from model organisms. But there is a clear need for a Web of Biology interface linking the resources and information of individual communities. Such an interface is critical to the needs of future biologists who will be tackling the biological problems that we consider intractable. Without such a centralized interface, the full range of capabilities of genes, RNA transcripts and proteins will not be appreciated. Time and effort will have been spent blindly creating databases that must be reworked at great expense to the public. Eventually, we're going to have to generate the necessary tools to give the bigger picture on biology. Preparing for it now is just good old-fashioned common sense.