It seems unlikely that supergenius Danny Hillis was unhappy with his time as a Disney ‘Imagineer’ . Where, according to this week’s Nature, he got to do very cool things like design themepark rides. But I wonder what Mickey thinks about Danny’s new job; killing a lot of mice with cancer.
Cancer research kills a lot of mice but in the following video Danny explains that he’s going to another level. It’s the result of a pinoneering approach that applies big data to bioscience. Together with oncologist David Agus, he’s bringing data science wizardry to proteomics, the study of the proteins associated with gene-expression.
First they built robots to automate and replicate blood protein analysis on level of unprecedented detail and precision. This prompted a holistic approach to solving cancer. “Cancer should be a verb”, argues Hillis because cancer is a process not a thing. The disease model of medicine works for problems caused by “things” like viruses or bacteria but it will break down for some like cancer. Cancers are more a kin to a process running out of control, so it doesn’t make sense to treat all brain cancers and say liver cancers as seperate groups based solely on where they occur. But to track the differences and the similarities you need data. a lot of it. So enter Danny’s army of tens of thousands of doomed mice. (Skip to 34:45 to hear about the dead mice.)
Nature Megadata: The odd couple