SAN DIEGO, Nov. 14 - Flops and bytes, not PCRs and mass spectrometers, will push the next wave of advances in genomics, according to Craig Venter.
Venter, delivering the opening address yesterday at the BioITWorld conference here, said that compute power will be the limiting factor in crunching, storing, and manipulating the gobs of data necessary for linking the promise of genomics to insights into gene function, protein interaction, and personalized medicine.
It's time to "get past the testimonial phase of what is supposed to be an objective science" he told GenomeWeb after the speech.
To get there, he said databases need to get to the exabyte range. His new mantra: "faster, cheaper computers."
To underscore his point, he said the Celera computers that sequenced the human genome--the 1.5 teraflop, 120 terabyte machines that took up 6,000 square feet of space--are relics.
"I wouldn't take it if they gave it to me," he said, speaking in his capacity as president of The Center for the Advancement of Genomics.
Venter predicted that by 2025, computers would reach 550 petaflops with 10 exabytes and occupy 50 square feet of floorspace. This kind of firepower would be needed to take the information from 30,000 genes, roughly 300,000 proteins, the "100 trillion ways" those proteins can interact over time, and tie it all to health care, he said.
Venter sees the computer chip, and not the hardware, as underlying these advances. "Intel and others [are] the driving force," he said.