A well-known researcher before age 30 and one of the youngest IEEE Fellows when he was 35, Mathukumalli Vidyasagar has spent the last 20 years working on turning his homeland, India, into a global engineering center.
The university professor left Canada for India in 1989 to focus on cutting-edge artificial intelligence, robotics, and industrial software. Now he’s in the United States to take on another challenge: overhauling costly drug development methods through early prediction of potential side effects.
In September he joined the mechanical engineering department of the University of Texas at Dallas as a holder of a Cecil and Ida Green Chair in Systems Biology Science. He also holds a joint appointment in the five-month-old biomedical engineering program at UT Dallas, a joint venture with the University of Texas Southwestern Medical Center at Dallas and the University of Texas at Arlington. Vidyasagar is conducting research into the interaction between computation and biology. He developed an interest in the topic when his daughter began studying cellular and molecular pathology at the University of Wisconsin, his alma mater.
Vidyasagar, 62, wants to come up with a set of mathematical models for how the majority of human beings are likely to react to a drug and how that drug might be altered to reduce side effects. Such a system could save hundreds of millions of dollars in development costs and reduce the side effects of new drugs. But to come up with such models, Vidyasagar first needs to determine what constitutes average physiology and how physiology varies from person to person.
“Advances in biological experimentation in the past two to three years have made it possible to get huge amounts of really good raw data, which we can turn into engineering models that can predict how certain populations will respond to various drugs before engaging in expensive clinical trials,” he says. “The vast majority of potential drugs cause too many unwanted side effects, but toxicity is not constant across people. The first step is to create a mathematical model of the average human and the variation across human beings from a cross section of people. We need to create models at various levels: cells, organs, and the whole body.”
One problem is that new drugs can perform well in animals but not in humans and most side effects tend to show up in the final steps of the drug development cycle, during human trials. The later the failure, the more costly it is to the pharmaceutical companies. Not only might they have to start over again, but they eventually pass on any extra costs through higher prices. Phase I trials, when healthy volunteers get an experimental drug, can run US $3 million to $5 million. Phase II, during which the drug is administered to patients with the targeted illness, costs $30 million to $50 million. Phase III trials, which determine the drug’s long-term toxicity, can run as high as $300 million.
“Why do humans sometimes react so differently from animals to certain drugs? The mathematical model we’re working on would let us see, before any trials, whether it’s necessary to tweak drug molecules in such a way that humans won’t get negative side effects,” Vidyasagar says.
Designing drugs to the maximum extent possible in a computer will be the next big research trend, he predicts. The cost of determining a person’s genome sequence is expected to fall from $60 000 to $5000 in the next year or two, encouraging correlations between an individual’s genetic makeup and responses to specific drugs.
An academic prodigy, Vidyasagar enrolled in the University of Missouri at age 13 in 1961, the year after his family moved from Tirupati, a small town in southeast India, to Columbia, Mo., when his father landed a job as a math professor at the university. The teen started in mechanical engineering but switched to electrical engineering after realizing he enjoyed tinkering with circuits. He transferred to the University of Wisconsin in 1964, graduating with a bachelor’s degree in 1965, a master’s in 1967, and a Ph.D. in 1969 at age 21.
For the next 20 years, he taught electrical engineering at Marquette University, in Milwaukee; Concordia University, in Montreal; and the University of Waterloo, Ont., Canada, before returning to India in 1989. There he served as director of the Ministry of Defence’s Centre for Artificial Intelligence and Robotics, in Bangalore. He went on to become executive vice president of Tata Consultancy Services, in Hyderabad, developing an industrial R&D lab focused on computational biology, qualitative finance, and e-security, before accepting the UT Dallas position.
“We’ll never be able to design new drugs completely by computer, but in 5 or 10 years we expect to be able to reduce the amount of trial and error involved in cutting down on side effects by 90 percent,” he says.