Entopsis: Molecular medicine meets machine-learning

That reads like a thinly veiled press release. I would like to know if it actually, you know, works. It seems to have no customers, which should cause raised eyebrows to the agiliests among us.

Also, regarding competitive methods, I could have sworn I heard exactly the same optimism regarding gene expression arrays. Like http://www.nature.com/scitable/topicpage/genetic-diagnosis-d... .

If the closest comparison is to mass spec then it feels like the choice was on technique - identification of certain fragments - rather than on outcome. If so, then that's a bit like saying a camel is a better way to get across the desert than a horse, omitting that cars and airplanes are likely better still.