Training for the Bioinformatics Boon

Call it a nerd's world. The explosion in genomic and proteomics research has ignited the demand for life scientists who can write computer code or make sense of the complex algorithms of genome sequencing research. The bioinformatics discipline preceded the genomic revolution, but because of the human genome research and its related studies, scientists with a yen for binary code can just about name their salaries at the world's burgeoning biotech companies. Feeling constricted by commerce? Big research institutes are opening or broadening bioinformatics departments and building information partnerships with their corporate colleagues. Duke University recently joined the computer-programming researchers club when it won a $7.2 million National Science Foundation grant to boost its bioinformatics resources. "There's a huge need for bioinformatics people," says Atul Butte, a fellow in bioinformatics at Children's Hospital in Boston and an instructor at Harvard University. "There are very few people who know exactly how to download the human genome right now. Even though it's publicly available, to do anything really fun with that requires some training and expertise; those people with expertise in bioinformatics are really hard to find." Even scientists who prefer a wet slide to a dry computer screen have confiscated the keyboards from the computer nerds and are trying their best to master the multiple databases and computer software programs required to find out what's known about a gene sequence or a microarray. In its June, 1999 report, the Biomedical Information Science and Technology Initiative, a National Institutes of Health working group noted that researchers who five years earlier avoided computers, now spend 90 percent of their time in front of their monitors. That report led to the creation of planning grants for the NIH Programs of Excellence in Biomedical Computing, a $10 million program to help institutions and scientists build their bioinformatics capabilities. The NSF also has made a push into computer-assisted research; its grants for Information Technology Research development across disciplines, including the life science, totaled $90 million in 2000, and the foundation expects to spend an additional $215 million in 2001, says Gary Strong, manager of the Information Technology Research Priority Area. To help scientists who want to deepen their expertise, Stanford University and the Marine Biology Laboratory in Woods Hole, Mass., offer short courses, ranging from three weeks to two years, and Duke University's Bioinformatics Shared Resource Lab offers three-day to weeklong courses in-house. Companies such as the Cambridge Mass.-based Lion Bioscience Research Inc., offers in-house training programs for their workers and clients. The company now plans to take the course on the road, marketing it on the Web. "Our main focus is really the bench scientists ... biologist-chemists who want to know more about bioinformatics," says Lion Bioscience CEO Reinhard Schneider. His parent company, Lion Bioscience AG, which was created in 1997 by six scientists working with the European Molecular Biology Laboratory, is working on a data integration project for Bayer. The courses offer Bayer scientists a bridge to computational biology. "It's not that [they] are bioinformatics experts, but they are certainly users who understand what [they] can know from informatics and what kind of tools are out there." Programming the Genomics Revolution Three years ago, genetics guru Eric Lander joined WorldWide Web pioneer Vincent Cerf at the White House to extol the benefits of bioinformatics. The scientists predicted the new discipline would revolutionize the life sciences and outlined the creation of a central online warehouse of genetic maps and data stored on supercomputers and made freely available on the World Wide Web. Biologists from all over the world could use these gene maps to augment their own research and enhance their ability to find new disease treatments, the two scientists predicted. "The placement of all that gene information in one database where it's accessible over the Net means that any researcher that finds a fragment of information can go to that database and find out if anyone else knows about that particular sequence, is this a pattern that has been seen before, and if so where and what does it mean," Cerf told the president and first lady at the "Millennium Evening at the White House," a videotaped discussion series. In the past three years, the replication of data on the Web has exceeded anyone's ability to centralize and streamline it, and institutions are looking for computer specialists who can build links between databases and find new ways to compare data. To help these multitalented scientists coordinate vast information stores and deliberate ethical issues, Greg Paris, executive director of the tools and technology group at Novartis Pharmaceuticals Corp. in Summit, NJ, is developing a special interest group of the American Society for Information Science and Technology. The group takes its definition of bioinformatics from a senior lecturer at the School of Informatics at City University in London, calling it "the generation, handling, storage, retrieval, and analysis of biological sequence data,"1 including nucleic acid sequences and proteins. Courtesy of Lucila Ohno-Machado Lucila Ohno-Machado The field requires more than handling data, however, stresses Lucila Ohno Machado, assistant professor of radiology and health science technology at Harvard University. Bioinformatics also requires intelligent decision-making and evaluation. "Even if you get access to the databases, how to use and what to use them for--knowing the research questions, knowing what can be answered--is important," says Ohno Machado. "There's a lot of hype. People need to know about the quality of the databases and what queries you can make." Training and Twinning Scientists Courtesy of Jim GoldenJim Golden To synthesize information, companies seek scientists with computational or statistical experience. Jim Golden, director of bioinformatics for CuraGen Corp. in New Haven, Conn., says that physics training has even proved helpful in understanding genes and evaluating the object and text data files used in the development of new medicines. "I think anyone that we hire with a background in software development has had some exposure to all those different data formats," says Golden, who originally studied mechanical engineering. "The challenge is internal, how we integrate. We spend a lot of developer time developing relationships between data types so that the computational biologists can make sense of the data." Golden says the marketplace often attracts scientists who combine university degrees in biology with self-training in computers and computation. In a competitive environment, pharmaceutical companies often cannot wait for their researchers to complete two-year programs. Schneider, from Lion, and Jason Johnson, a senior research scientist in computational biology at the Cambridge, Mass., Pfizer Discovery Technology Center, often hire effective computational people and train them in biology. "It's easier to start as a computer scientist and learn the biology because the biology is changing so rapidly that you have to learn the new stuff anyway," Johnson says. "You have to be up-to-date and the newest stuff is the most important anyway." But not all companies rally around the self-trained scientists or the computer specialists functioning as biologists in-training. "If we are talking only of people who interpret biology, they should have dual backgrounds to understand it well and to develop methods to interpret biological phenomenon," says Thodoros Topaloglou, director of data at Incyte Genomics in Palo Alto, Calif. "We need people who can take data from this point on and develop a product." To supply the profession with dual-trained scientists, the University of Pittsburgh has eschewed short training courses in favor of full two-year pre- and post-doctorate degree programs for computational biologists. "They need to support the biologists in a way that they can adapt the tools for the non-routine," says Charles Friedman, associate vice chancellor for biomedical informatics at the University of Pittsburgh. "We need to figure out what bioinformaticians are and how scientists should be trained so they can effectively do the two roles and build the next generation of tools." Paula Park can be contacted at ppark@the-scientist.com 1. N.J. Cole, D. Bawden, "Bioinformatics in the pharmaceutical industry," Journal of Documentation, 52[1]:51-68, 1996. Bioinformatician Starters At Duke Courtesy of Simon LinSimon Lin Duke University offers crash courses for people trying to get first-class seats on the bioinformatics bandwagon. The two- to three-day seminars provide biologists opportunities to learn a little computing, and give computer software experts a view into life science disciplines. Simon Lin, class organizer at Duke's bioinformatics shared research program, says the courses bridge the dichotomy of bioinformatics, which he attributes to the shortage of scientists with both computing and research experience. The university also has launched new graduate courses, and plans a full PhD program within the next year. "There is not a lot of coursework available for training in this," Lin says. "So many people are eager to take advantage of the seminars." The workshops specialize in bioinformatics programs such as BioPro and programming languages such as Java. The courses also cover biology for statisticians, Beta-base design, and sequence analysis; they cost $300 to $500 a day. Lin says that the school is planning a certificate program of five to six workshops for entry-level bioinformaticians. Duke also offers five separate graduate bioinformatics courses, each conducted over several weeks. Select students from the program qualify for the Bioinformatics Shared Resources Scholar Program--four to eight weeks of hands-on training with the university's many computer resources. Rhazes Spell, a graduate biomedical engineering student, says, "The courses are a good blend, covering a broad range of topics. We get to work with a lot of tools and techniques." Carla Rydholm, a graduate student in genetics, expresses satisfaction in venturing outside her discipline and into the bioinformatics courses. "If anything, I have become more flexible as a scientist," she says. "Someone like myself with a biology background finds a whole new way to look at something when being lectured by a statistician." Short seminars and graduate school courses can help scientists warm the computer keys, but they may not instantly create a cadre of bioinformatics specialists, says Dave Jensen, managing director of Search Masters International in Sedona, Ariz. "Maybe two years ago, an individual trained in another area could step into the job market with only a crash course under his belt," he observes. "But that is not the case at this time." Duke's future PhD program in bioinformatics, Lin predicts, will enable students to integrate computational and biological science expertise. "Bioinformatics will be a skill that the next generation of the biologists will possess," Lin says. Stanford and Harvard universities, Massachusetts Institute of Technology, and the University of Washington, among others, already offer PhD programs in bioinformatics. "People tend to perk up when you say you're studying bioinformatics," Spell says, adding that he believes he may find job openings in some of the bioinformatics specialties he studied at Duke. "I feel the skill sets that we learn are very transferable." --Hal Cohen

Training for the Bioinformatics Boon

Interested in reading more?

Become a Member of