Someone (I can’t recall the source) once said that rather than envy the creative power of nature, we ought to emulate it. Genetic algorithms have, for a couple of decades now, allowed scientists and engineers to harness the power of evolutionary processes to evolve (literally) new solutions to hard problems in silico. A new technology developed by George Church at Harvard seems aimed at allowing scientists to do the same thing in vitro.

Here’s the story, by Arielle Fridson, writing for Bloomberg: Harvard Scientists Speed Up Evolution, Aim to Sell Technology to DuPont

Evolution often plays out over millennia. George Church says he can make it happen in days.

He and his team of researchers at Harvard Medical School in Boston invented a table-top machine that can generate multiple changes in the DNA of bacteria all at once, speeding up the evolutionary process. Aiming to sell the technology for use in chemicals, energy, and the drug industry, they’re talking with DuPont Co. and other companies, members of the team said….

This of course is pretty much the exact opposite of the kind of genetic engineering that makes use of careful insertion of genes with known functions. It allows a kind of generate-and-test procedure: generate a bunch of mutations, and see which ones do what you wanted them to do. This allows for enormous creativity, as the literature on genetic algorithms attests. (Back in grad school, I fiddled with G.A.’s and ecological programming a bit, and got to watch novel — and sometimes radically weird — solutions to hard problems literally evolve.) Along with this creativity also comes risk, when the things evolving are potential pathogens. Church is right to worry (as the Bloomberg story notes he does) about the potential for misuse.