Image: Erica P. Johnson
Bioinformatics is growing up. Science's hottest information tool is coming into adolescence and has transformed the way research is conducted. For life scientists, this transformation means that those who lack skills to integrate informatics in their work are in danger of not staying competitive and also of not staying employed.
"There is no doubt in my mind that the life sciences have already turned very informational, and once the train is on the track you can't reverse it," says Mark Hall, program director of life science research at International Data, a provider of technology intelligence, industry analysis, and market data. "If you are going to break through in new levels of biology, it is going to be assisted by the use of information. It is going to have to be done on math scales and [by looking] at problems the way technology does."
ACADEMIC RESISTANCE While the life sciences have embraced bioinformatics, not every life scientist has. Hall and others say that on the whole, US universities lack the teachers for needed bioinformatics courses. Until universities change, he says, departments should take a multidisciplinary approach to instruction.
"There is an enormous amount of resistance from established faculty," says J.W. Bizzaro, founder and president of Bioinformatics.org, a nonprofit organization based in Hudson, Mass. "Biologists have no real respect for theorists and below that [for] the 'computationists.' It isn't respected or understood." Still, the need for people qualified in bioinformatics, evident in the advertisements in the science journals, has kick-started degreed bioinformatics programs at some universities. At latest count, 45 schools have formal programs, ranging from bachelor's degree to doctorate, and new ones are added each year.1 "Courses are being set up by universities to teach the necessary biology to math people and the needed math to biologists," says W. Graham Richards, chairman of the Department of Chemistry at Oxford University. "People are moving into the field with extra skills."
Others in graduate school and postdoctoral programs in the life sciences are gaining the skills, because they have mentors who aren't frozen in the amber of traditional biology.
They encourage their charges to delve into computer and math courses. "The best [mentors] are trying to train people differently from the way they were trained," says Russ Altman, an associate professor of genetics and medicine (and computer science by courtesy) at Stanford University. "The best biologists are encouraging students to take some math during graduate school training ... in the first two years to become familiar with this kind of thinking. Biologists will need to understand the people they are talking to and have at least an introduction to their field."
The important role of bioinformatics seems a foregone conclusion. If science is going to take quantum leaps, scientists must find ways to efficiently mine the rich veins of information hidden inside the compiled data, Hall and others say. "There is an increasing interest from professional computer scientists and really serious mathematicians," Richards says, "not merely people who know the biology but people looking in a real hard science way. That is growing."
So what new skills will postdocs need to ensure that they don't become science relics? The answer is math, statistics, and knowledge of a scripting language for computers. "Computers will become ubiquitous in biology," Bizzaro says. "Knowledge of data bases is important, as is a good knowledge of statistics."
GRID ART Until recently, Richards says, gene sequencing had been the main target of bioinformatics. But now scientists are using the tool to hunt down disease therapies. Richards currently employs bioinformatics in grid computing.2 He has managed to recruit 1.74 million people in 215 countries, who allow him to use the screensaver time on their personal computers to sort through cancer drugs. Using a single software platform, the grid effectively strings together all the computers and makes them into one virtual computer. Grid computing provides a massive amount of computing power at very little cost. "The hardware limitation is essentially a thing of the past or promises to be," says Richards. "You can ... combine data in a numerical form, but also you can have images like lungs or visuals or graphic data."
Grid computing is but one area where bioinformatics can be applied. Altman believes that the real opportunities in the next decade will come to scientists who can understand a system and develop modeling and simulating pathways with bioinformatics. "If I were a young student entering the field now, I would make sure I understood how to run simulations," he says.
As people with math and computer backgrounds expand the bioinformatics field to fit their own expertise, young biologists will need to know their specialties inside and out, so that they can communicate and work with mathematicians and computer scientists. The trickiest part will be making sure everyone understands what the other is talking about. To that end, Altman says, life scientists will need to learn to speak and think in different ways. Getting the language is very important. "The future of biology is going to be very 'mathy' and 'computery,'" says Altman, creating a few new terms of his own. "People who went into biology because they don't like math and computer science may have a problem."
Altman says it is equally important for biologists to take computer courses, if for no other reason than to learn the way computer technology people think. Biologists, he says, are organic and are comfortable with ambiguity. Computer people are not. "If you are familiar [with] looking at data, you have a clear appreciation for the comparison and analysis that is done, and where it's weak and where it's strong," he says. "People who can do that and are good experimentalists as well will be the cream of the crop."
No doubt the cream of the crop will get the best jobs. But it won't be long before biotech and pharmaceutical companies will expect every scientist to possess computer expertise as well as their domain expertise. How they will use bioinformatics will depend on each company. "Some pharmaceutical companies aren't necessarily all that crazy about bioinformatics," Hall says. "They think it will help but [that] it isn't the answer. Others are much more information oriented and value the use of information technology to solve problems. They will push their people to adopt these technologies."
Companies are already hunting for highly skilled scientists. CuraGen, a New Haven, Conn., genomics-based company that combines engineering principles with biology and information technology to develop new drug therapies, puts together teams of people from different fields. "You try to bring together people with knowledge of one domain as well as ... computer [science] and tie them together as a team," says John Murphy, the company's vice president and chief information officer.
Murphy says that future postdocs will be expected to have a deeper understanding of bioinformatics. They will need to know software development and have some expertise in computer programming. Murphy already sees changes in the way some graduate students and postdocs are being trained, compared with the training of 20 years ago.
As they learn to approach problems differently, he says, they will become more productive. "The best of the best will be a person who combines an in-depth knowledge of an area of biology or medicine and who is also able to look at things as a generalist in a broad overview," Murphy adds. "You'll need to be able to think about the big picture and how your piece fits in."
Bob Calandra (firstname.lastname@example.org) is a freelance writer in Philadelphia.