James Kent, a graduate student in the laboratory of Alan Zahler at the University of California at Santa Cruz, recently was accorded hero status by the New York Times for his role in the completion of the sequencing of the human genome. In just four weeks, Kent produced a computer program known as the GigAssembler that enabled the public consortium of sequencing laboratories to splice together 400,000 overlapping DNA fragments of the human genome. "I felt proud and a little embarrassed," said Kent when asked about the Times story. "There are a lot of other people working very hard on this project."

Kent was being modest. To complete the GigAssembler, he wrote code night and day at a pace so furious he had to ice his painful wrists due to flare-ups of repetitive-stress injury. "The month of March 20 to June 20, 2000, when I put together the first working version of the GigAssembler, does rank as one of the heaviest programming months of my life, but probably not the heaviest," said Kent, who at 41 had had a career in computer graphics before embarking on a graduate program in biology. "I figured they'd need somebody to index and hyperlink the human genome soon enough, so I decided to switch to biology."

Kent's prediction proved uncannily accurate, and he rose to the challenge. The public consortium hadn't realized how much it needed such an assembly program until the last minute, reported the Times, when competition heated up between it and Celera Genomics, the private Maryland biotechnology company that threatened to run away with the sequencing crown. Kent managed to create a program that allowed the public consortium, under the direction of the National Institutes of Health's National Human Genome Research Institute, to remain competitive with Celera. As a result, the two groups announced their completion of the sequencing of the human genome on the same day - June 26, 2000.

Kent's story illustrates beautifully the impact that the new field of bioinformatics is having on biological research. The complexity and vast quantity of genomic data that have become available over the past few years have made it necessary to develop new computational methods to assimilate and utilize this information. "Without bioinformatics, there would be no genomics," said Charles DeLisi, professor of biomedical engineering, dean of the College of Engineering, and founder of the Bioinformatics Graduate Program at Boston University. This past January 8, DeLisi received a President's Citizens Medal for his contributions to the field of genomics.

Bioinformatics has emerged as a distinct discipline that straddles the interface between the traditional biological sciences and the computer sciences and advanced computational methodologies. It is rapidly becoming a powerful new approach to understanding life, and it may well reverse the reductionist paradigm that has held sway in molecular biology ever since Erwin Schrodinger turned on a generation of physicists to biology with the publication of What is Life? more than 50 years ago.

Over the past half century, molecular biologists have focused on understanding the "parts list" for living organisms. A researcher would devote enormous time and effort to isolating and characterizing a particular gene, enzyme, or other protein, often spending an entire career studying one or a handful of macromolecules. Sometimes such research would reveal the interaction of a protein with one or a few other proteins or genes, thereby providing a glimpse of the complex molecular matrix of life. However, even the most talented - or luckiest - researcher could hope to gain only the narrowest view of the enormous interacting network of genes, proteins, and regulatory molecules that are found in even the simplest living organisms.

The new discipline of bioinformatics promises to provide the tools needed to attack the complexity of conducting holistic biological research. "Because of this complexity, biology will eventually become the most computational science, surpassing physics," said DeLisi, who predicts that within the next 10 to 15 years bioinformatics will become an integral part of biology.

With the maturation of genomics, it is now possible to go to a database and find nucleotide sequence data for entire genomes as well as the deduced or experimentally determined amino acid sequence of many of the proteins they encode. Complementing this vast store of information are new technologies such as gene chips that can reveal in a single experiment the pattern of gene expression of an entire genome in a particular cell or tissue type, a capability undreamed of just a few years ago. Such developments are rapidly transforming biological research, allowing investigators to follow the ripples of perturbations as they traverse the molecular matrix during normal cellular function as well as in dysfunction. Bioinformatics promises to provide the tools necessary to exploit these enormously exciting developments. In the future, some degree of facility with the basic techniques of this rapidly developing discipline is likely to be a prerequisite for success in many areas of biology.

Pharmaceutical research will clearly be one major benefactor of developments in bioinformatics. Already, through the use of computational techniques to search for genes similar to those known to encode proteins on which existing drugs act, hundreds of potential new drug targets have been identified. In the future, virtual toxicology screening may be the first step in predicting the effects of new chemicals on complex metabolic pathways. In addition, bioinformatics will likely provide the methodology finally to make highly accurate predictions about protein tertiary structure based on amino acid sequences and a viable means to design drugs based on computer simulation of the docking of small molecules to the predicted protein architecture.

One of the major surprises that emerged from the completion of the human genome is that the number of genes is somewhere between 30,000 and 40,000 - much smaller than the 100,000 or more that many investigators had anticipated. This raises an obvious question: How is the rich diversity of protein structures generated from such a small number of genes? The answer appears to include the fact that protein domains are mixed and matched from one protein to another to make a much broader array of protein structures and functions than would be expected from a genome containing only 30,000 to 40,000 genes. New computational techniques are being developed to identify these distinct protein domains and to understand how they are combined to produce a large repertoire of unique proteins.

Bioinformatics is not limited to understanding DNA and amino acid sequence databases. New computational methods will likely transform taxonomic and phylogenic studies as well as our ability to understand and predict the results of complex signal transduction cascades and the kinetics of intricate metabolic pathways. The Texas A&M University Working Group in Bioinformatics focuses on using new computational methodologies to access information in botanical databases and developing new approaches for the expression of biodiversity data. In short, the new discipline of bioinformatics appears destined to transform all areas of biological research.

Boston University (BU) was one of the first schools to develop a graduate program in bioinformatics. Its bioinformatics master's and doctoral programs accept applicants with undergraduate backgrounds in either the biological sciences or computer science and mathematics, said DeLisi; there are currently 32 doctoral students. Those with undergraduate degrees in the biological sciences tend to concentrate on graduate courses in the quantitative and computer science fields, while students with computer science and math backgrounds concentrate on courses in biochemistry, molecular biology, and cell biology. A unique aspect of these programs is the selection of two research advisers to direct the student's research; one will have expertise in computational areas, the other in the chemical or biological sciences.

The BU program also includes internships, which allow students to gain industrial experience as part of their graduate training. The internship can range from participating in "grand rounds" to gain an overview of the field from an industrial perspective, to spending three to nine months on-site doing industrial research with a particular company. Some 35 to 45 companies participate in a senior project day.

Other universities with graduate programs in bioinformatics include the University of Michigan, the University of California at Los Angeles, the University of California at San Diego, North Carolina State University, the University of Toronto, and George Mason University. In addition, many biology, biochemistry, and molecular biology graduate programs now include bioinformatics courses and research opportunities. It's likely that in the near future many more universities will develop specific programs in bioinformatics.

Competition for the available graduate programs is intense. For example, Boston University selected its most recent class of 10 students from more than 300 applicants, noted DeLisi, and most of those 10 had some kind of industry experience before entering the program. Moreover, the course of study is very rigorous, requiring expertise in both biology and computer science. If you are interested in graduate study in bioinformatics, says David Haussler, professor of computer science at the University of California at Santa Cruz and a Howard Hughes Medical Institute investigator, "take computer science courses so you can understand how algorithms are designed and implement them well. Take statistics, linear algebra, discrete mathematics, and differential equations. Take genetics, cell biology, molecular biology, organic chemistry, and biochemistry. Don't plan on an extensive social life!"