In the early days of scientific discovery, biologists had to be skilled in disciplines outside of their own biological realm, such as in the art of blowing glassware. Similarly, young biologists today will have to learn and apply techniques from other disciplines and embrace new ways of thinking if they are to pursue and advance careers in bioinformatics.

Just how researchers will be affected by the growing overlap between biological and technological disciplines was discussed last Sunday, 20 February at the American Association for the Advancement of Science annual meeting in Washington, D.C. The New Careers in Biology: Building Biodiversity Informatics for the Millennium panel says that interdisciplinary science is critical in understanding the complexity of living organisms and ecosystems. To properly value and advance scientific knowledge, data from thousands of sources will have to be standardized, pooled, and reevaluated. And as bioinformatic disciplines develop, scientists who can manage computer databases and conduct scientific experiments will be highly sought after. Ensuring young researchers are trained well is essential and the panel calls for more initiatives and federal investment to successfully unite seemingly disparate disciplines.

Robert Robbins, a panel member and vice president of information technology at the Fred Hutchinson Cancer Research Center, said that bioinformatics is fast becoming an independent discipline and "not just an adjunct to biology." He described the evolving science as "magic technology" and predicted that bioinformatics will become so advanced that the distinction between living and nonliving will be blurred. Citing Moore's Law--which states that every 18 months the number of transistors on a silicon chip doubles--Robbins explained that "the volume of electronic data will be enormous." The universal availability and storage of this data will change the way investigators do research, creating new opportunities for researchers willing to bridge information technology and life science. "Like physics needs calculus, biology needs bioinformatics," he added. Industry has been fast to capitalize on the needs of this growing market and more companies are churning out genetic data. "In the last 10 weeks, more sequences were added to the GenBank database than in the first 10 years of genomics," reveals Robbins. "The real question," for researchers, he points out, "is understanding all this data."

In the future, according to Robbins, biological research is going to move from being hypothesis-driven to being data-driven. Researchers will need to be proficient in library science and management science to fully exploit careers in bioinformatics, Robbins says.

In 20 years, it won't be necessary to employ bioinformaticians as specialized faculty within academia, believes Robbins. "Informatics will be part of their everyday skills," he says. Biologists must learn to "present their problems in a sufficiently abstract way," so that these emerging technologies can be successfully applied.

One such application is to advance biodiversity research. The U.S. Geological Survey-led National Biological Information Infrastructure (NBII), for example, is developing a "distributed electronic biodiversity and environment 'database,' " to allow researchers to pool and compare vast amounts of information. There are two practical obstacles: Creating the database resources and recruiting qualified individuals. Biological "metadata standards" will be needed to describe all data from all sources in the same format. NBII hopes to collate information from natural history museums for example, but collecting and storing those details into a common database is complicated. "How can we digitally capture that data? It's all untapped because the data is not electronic," clarifies fellow panelist and NBII executive secretary Anne Frondorf. "Species names are key to data indexing," says Frondorf. "But the same species can have different names. We need to have people who have an understanding of biological issues as well as computer science," she stresses.

So what is the scientific community doing to help promote interest in these areas? Carter Kimsey, program manager for postdoctoral research fellowships in the National Science Foundation's (NSF's) Division of Biological Infrastructure, spoke of training opportunities in informatics. "The postdoc period is the most appropriate stage for cross-training," she believes. NSF fellowships are available for cross-training and Kimsey says 65 candidates applied for the 18 awards made last year. James Edwards, deputy assistant director for NSF's Directorate for Biological Sciences, adds that "all NSF fellowships have an [optional] international component" to spend time overseas doing research. Only one of last year's fellows opted to do that however. Encouraging research abroad is vital, Edwards says, because other countries are making significant progress in bioinformatics. "Look at Australia--they're way ahead of the U.S."

Other training initiatives include designing specific bioinformatics curriculae for higher education courses. Arik Melikyan, a visiting math professor at Georgia Institute of Technology, spoke of the master's degree in bioinformatics that was created there. Supported by the Sloan Foundation, the program began in Fall 1999, introducing students to the informatic concepts and math. But the program does not culminate in a written thesis and there is no research requirement. A further criticism is that the prerequisites for entry (a GRE score of at least 1800) may be too stringent, diminishing the interest of many young scientists while artificially defining the new population of bioinformaticists.

But training is necessary: "Modern biology is going to need a much bigger dose of mathematics," says Edwards. Biological researchers should prepare to learn "at least introductory calculus, statistics, and graph theory." Modeling and simulation are also becoming more and more valuable, especially in drug design. "We need people who are literate in these fields," reemphasizes Edwards. The advent of bioinformatics may also improve the current postdoc-in-limbo situation. Not only will trained young researchers be "snapped up," but no panel member expects a bioinformatics postdoc to serve more than one fellowship before being hired. They all agree that industry is expected to do most of that recruiting.

Bonnie Carroll from the U.S. Geological Survey, who moderated the session, closed the discussion by appealing to those interested to create a forum in which to flesh out the problems and "bring faculty and students together." The magic behind the technology may soon be revealed to all.