Polly Want a Genome

When Duke neurobiologist Erich Jarvis started trying to decode the genes believed to control a parrot’s ability to imitate its owner, he quickly ran into a problem. Typically, scientists assemble genomes in a process like building a jigsaw puzzle, using sequencing machines to crank out small segments of DNA code and then figuring out how to piece them together into a coherent sequence. But the regions that interested Jarvis—involving complex genes regulating brain circuitry—proved too complicated to stitch together from fragments.

To sort it all out, Jarvis and a team of collaborators combined two sequencing technologies— one that reads longer segments of code but is prone to mistakes, and one that produces shorter, more reliable sequences. With the longer read as their rough draft, the team used shorter segments to proof the code and correct errors. They tested the technique on known genetic sequences in corn and bacteria to confirm its reliability.

The hybrid approach helped Jarvis produce the first detailed map of the genes that make parrots parrot—data that could also explain how similar speech-development genes in humans function, he notes. And it may well aid biologists in trying to sort out other knotty chapters in the genetic storybook, like the complex genes involved in cancer.