Thursday, April 9, 2009

Startup #2


This idea is slightly outlandish.

Concept: use a genetic sequencing machine to adequately match a potential child with adopted parents.

Let's set aside the legal and ethical ramifications of such a concept for the moment. To get started you would need a sequencing machine, and a clean room (aka lab space) to run genetic material through the machine. In 2008, the cheapest machine was about $150,000 and could do 10 billion base pairs in an 80 hour run.

Basically, the possibility now exists to apply sequencing techniques to personalized medicine. But why not apply it to other areas? (like adoption matching). From the aforementioned article:

"An inexpensive new gene-sequencing machine is due to hit the market next month, and its creators hope that it will make sequencing more common, ultimately giving a boost to personalized medicine. The machine is the brainchild of George Church, a genomics pioneer who developed the first direct sequencing technology as a graduate student in the 1980s and helped initiate the Human Genome Project soon after.

Church sees greater access to sequencing as a vital component in the drive toward personalized medicine, in which treatments and preventative medicine are tailored to an individual's genetic makeup. The new machine, which was developed with an "open source" philosophy, was commercialized by Danaher Motion, based in Salem, NH, with the specific intent of keeping costs low. "It seems like the biomedical-instrument field in general tries not to commoditize," says Church, who heads the Center for Computational Genetics at Harvard Medical School, in Boston. "It tries to keep profit margins high and slow the inevitable decrease in cost." The Danaher device will cost roughly $150,000, a third to a tenth of the cost of systems currently on the market."

So to do this you're going to need some people with scientific background to know which sequence encodes certain genes leading to certain phenotypes. Then you would need someone with the bioinformatics skills to assign probabilities to the child turning out a certain way (the way the potential adopted parents wish).

This is something I definitely think will be developed, or at the very least get some serious VC $. 23andMe would have seemed impossible 5 years ago.