simply brilliant....,

The Scientist | A project at Harvard Medical School aims to bring music to medicine in a way that goes beyond setting the mood in the waiting room. Gene transcription and translation are anything but simple. But by combining modern statistics with the sounds of a sweet melody, bioinformatician Gil Alterovitz may make interpreting these complex phenomena and diagnosing the diseases that result from abnormalities in gene expression much more manageable tasks.

"I think it's brilliant that Gil is using a completely different channel for communicating complex genomic information," Latin and ballroom DJ Taro Muso writes in an email to The Scientist. "I've always wondered why doctors don't seem to use their ears beyond listening for natural bodily sounds."

"It's deceptively simple," says bioinformatician Yves Lussier of the University of Chicago. "It was conceptually challenging to come up with it, but once we know of it, it's obvious we should have tried that in addition to visualization techniques we have been using."

By boiling down gene expression data to just a few components -- variables that condense one or more parameters of data -- and assigning each of those components a different note and musical instrument, Alterovitz and his colleagues are literally making genetics musical.

The team carefully chooses the notes such that normal gene expression patterns sound pleasantly in tune, while abnormal data yield discordant sounds. "When you hear inharmonious music it kind of catches your attention," Alterovitz says, "and that would be a sign of a pathological problem."

"Even amateur musicians can tell the difference between various chords," Muso agrees, "so there is a definite potential for motivated biologists to use harmony as a screening method."

Alterovitz got the idea ten years ago while doing his PhD at the Massachusetts Institute of Technology. When he donned his scrubs and joined surgeons in the operating room as part of his graduate research project, he was distracted by the numerous monitors measuring nearly two dozen biological signals. Sometimes an alarm would go off, he recalls, but most of the time it wasn't really relevant, and they were simply turned off and ignored.

"Wouldn't it be useful if we somehow integrated [all] those variables so that we could present something that was not just a binary alarm but holistic information about the whole system?"