IBM's choice of protein folding as the focus of its project to develop the supercomputer Blue Gene is part of a strategy to position itself as a leader in computational biology, as the post-sequencing phase of the Human Genome Project approaches (see above).
The company has regrouped its life-science activities into a single unit focused on genomics and drug design. Sharon Nunes, senior research manager at IBM's Computational Biology Center at the company's Deep Computing Institute in New York, says that predicting protein structure will be a key use of genome data. “Using today's computers it would take a thousand years to fold a protein,” she says.
“Blue Gene is tremendously exciting,” says Christopher Dobson, who works on protein folding at the University of Oxford. “Folding is how the cell converts information from DNA into biological function, how the cell targets proteins to particular parts of the cell, controls and localizes them, and is key to turning activity on and off,” he says.
Some are sceptical. Computing power alone is not enough, says Geoff Barton, a scientist at the European Bioinformatics Institute in Cambridge. He argues that solving folding will require enormous progress in understanding the interactions between the atoms in a protein.
“There is no guarantee that if you find the minimum energy conformation it will actually be the native protein conformation, because the potentials used to model the pairwise interactions are imperfect. There are also difficulties with modelling the true solvent environment of the protein.”
Ajay Royyuru, a researcher at IBM's Computational Biology Center, says this will be a major thrust of the Blue Gene project. Dobson says the timing is right in that “many of the fundamental principles of folding are beginning to emerge”.
Whatever the reservations about protein folding, few biologists doubt that, if it succeeds, the project's contribution to handling the mountains of data emerging from genome sequencing programmes will be revolutionary.