A federal advisory panel recommended last week that the National Institutes of Health invest heavily in computer technology and training for government-supported biomedical researchers, saying the coming rush of scientific data--and competition from private industry--require this enhanced expertise.

"We're not keeping up," said David Botstein, chairman of the Department of Genetics at the Stanford University Medical School and the head of the panel. "It's not enough to teach computer scientists biology--we must teach biologists computer science."

Pharmaceutical and biotechnology companies already have turned to computers to analyze information on new treatments and to better understand the flood of information emerging from current efforts to decipher the human body's genetic makeup.

The recommendations stressed that federally funded efforts must now do the same and that they should encompass virtually all aspects of medicine--especially biology and genetics--or risk losing an important edge in scientific advances.

The recommendations almost certainly will be embraced by NIH Director Harold Varmus, who already has indicated his support for beefing up biomedical computing.

He and others have said that many of today's established scientists lack this technical knowledge. While many researchers, particularly younger ones, are becoming computer savvy, "it is in the interest of the NIH to accelerate the process," the panel's report said.

The centerpiece of the advisory group's recommendations is a call for NIH to establish up to 20 training centers to teach computer-based medicine.

With these centers, "the best opportunities can be created for doing and learning at the interfaces among biology, mathematics and computation," the report said.