The sequencing of the human genome ensures that cancer can't hide its secrets forever. However, like looking for specific specks of sand at the beach, finding each secret demands sifting through a mind-boggling — and time-consuming — expanse of molecular variables.
The enormous challenge has given rise to a new field of science: computational biology. The Hutchinson Center has established computational biology — along with the related field of bioinformatics — as the newest of its 18 research programs. Dr. Robert Gentleman, who was recruited last year, leads the Herbold Computational Biology Program.
By using technology to marry the disciplines of biology and mathematics or statistics, the program promises to improve the productivity, speed, accuracy and cost-effectiveness of cancer-research experiments. This will revolutionize the world's understanding of cancer and accelerate the flow of lifesaving discoveries. "We have to move in this direction if we want to be successful," Gentleman said.
Computational biology defined
Before coming to the Hutchinson Center last December, Gentleman was an associate professor in the Department of Biostatistical Science at Dana-Farber Cancer Institute and Harvard University. He has authored 30 software packages for statistical analyses and computational biology including R, a widely used language and suite of software for statistical computing.
Bioinformatics and computational biology attack the same problem — how to process the oceans of data involved in searching for the molecular clues to cancer — but from different directions. Bioinformatics involves developing tools to analyze the large sets of data generated by traditional or "wet-lab" experiments. Computational biology involves conducting experiments "in silico" — using powerful computers and sophisticated software and mathematical modeling to perform virtual experiments based on data that already exists.
Just as a flight simulator is no substitute for actually flying, computational biology is not intended to replace laboratory research. Rather, the goal is to make laboratory research more efficient by eliminating numerous preliminary steps. "Computational biology offers a way for us to learn enough about a problem to guide decisions about the most critical experiments to perform in the wet lab," Gentleman said.
Plans for the program
For example, even if experimenting with simple yeast cells, a scientist seeking insight into the combinations of genetic mutations that can lead to cancer would need to create and analyze millions of different strains of yeast. "A better — and certainly more economical approach — would be for the lab scientist to set up a smaller-scale experiment to provide some preliminary results for a computational biologist to work with," Gentleman said. "That information would be used to test the approach on a larger scale using a computer, which in turn would generate data about which combinations or how many might be the most important to focus on in the lab."
The Herbold Computational Biology Program will support numerous research endeavors ranging from DNA analysis to systems biology to disease modeling. As additional faculty join the program, they not only will conduct research of their own, but also will collaborate closely with members of the Center's laboratory divisions in a rising tide of interdisciplinary research. "We hope that this program will catalyze many new and productive research interactions," Gentleman said.