The National Institutes of Health should invest $100 million to buy a stateoftheart supercomputer and another $100 million to develop software and train scientists to use the system for biomedical research, the head of a pharmaceutical company told Congress last week. Frederick Hausheer, a physi
The National Institutes of Health should invest $100 million to buy a state-of-the-art supercomputer and another $100 million to develop software and train scientists to use the system for biomedical research, the head of a pharmaceutical company told Congress last week.
Frederick Hausheer, a physician whose San Antonio company, BioNumerik Pharmaceuticals Inc., makes cancer drugs, said NIH needs a larger share of the budget from the Clinton administration's new Information Technology for the 21st Century Initiative (IT-squared).
NIH's "use of supercomputing and information technology'' and its "absolute computing power'' is "far behind the capabilities'' of the other five agencies that would get most of the proposed $366 million IT-squared budget next year, Hausheer said. He added that more powerful supercomputers are necessary to model the effects of potential new drugs on "one of the most complicated entities on Earth - the human body.''
But Rep. Nick Smith (R-Mich.), chairman of the House Science Committee's Basic Research Subcommittee, which held the March 16 hearing, said Congress will have to set priorities for all of the government's research programs. "We will continue to be subject to constraints on overall spending,'' he said.
The subcommittee is working on a bill to authorize IT-squared. Smith said after the hearing that he is considering whether the new program should be merged with the $148 million Computing, Information and Communications (CIC) program, which includes most of the same agencies.
IT-squared is the result of recommendations by the President's Information Technology Advisory Committee (PITAC), which suggested that the government spend more money on "fundamental'' computer science research. Existing programs, such as CIC, mainly support projects that address agencies' mission-specific problems.
Under the IT-squared plan, NIH would receive $6 million, the same as the National Oceanic and Atmospheric Administration. Of the remainder, $148 million would go to the National Science Foundation, $100 million to the Defense Department, $70 million to the Energy Department and $38 million to NASA.
Ken Kennedy, a Rice University professor and co-chairman of PITAC, said the panel's recommendations assumed current programs would continue. "I'm one of those who believes all fundamental research needs to be increased,'' he said, adding that although IT is essential to all scientific research, it receives one in 75 federal research dollars.
Stephen Wolff, the former director of networking and communications research at NSF who now runs Cisco Systems Inc.'s Advanced Internet Division, said the IT-squared plan should place more emphasis on cryptography and privacy research. "It is somewhat disappointing that research on cryptography receives only casual mention [in the plan] and research on the technology of privacy no mention at all.''
More research in "the basic science and mathematics of cryptography and privacy is needed,'' he said, to build technology that meets "the legitimate needs of law enforcement officials and the legitimate desires for individual privacy.''
Regarding NIH funding, Neal Lane, President Clinton's science adviser, said he expects NIH "will be a much larger participant'' in IT-squared after the agency reviews its computing needs. Last fall, NIH Director Harold Varmus convened a 15-member advisory committee to review the agency's biomedical computing requirements. An NIH spokesman said the group is expected to deliver its recommendations next month.
"Everyone agrees we need to move forward with information technology [research],'' Smith said, but he questions how best to structure these programs.
NEXT STORY: OMB report: Feds must shore up state Y2K efforts