Your whole genome for $5000

In yesterday’s NY Times:

The cost of determining a persons complete genetic blueprint is about to plummet again to $5,000.
That is the price that a start-up company called Complete Genomics says it will start charging next year for determining the sequence of the genetic code that makes up the DNA in one set of human chromosomes. The company is set to announce its plans on Monday.
[T]he cost of DNA sequencing has dropped by a factor of 10 every year for the last four years, a faster rate of decline than even for computers, Dr. [George] Church said.

That’s just incredible. I suppose if you just paid for Knome to do it for $350,000, you’d feel pretty burned…

Knome, a company that offers to provide consumers with their DNA sequence, charges $350,000. But that is a retail price that includes not just the sequencing costs but also the analysis of the data and the customer service.
Complete Genomics will not offer a service to consumers. But it will provide sequencing for consumer-oriented companies like Knome.
Knome is already exploring farming out its sequencing to Complete Genomics. We anticipate wed be able to significantly drop our price, said Jorge C. Conde, the chief executive of Knome, which is based in Cambridge, Mass.

Heh. What are “the analysis of the data and the customer service” worth? I mean, since the SNP-handling companies like 23andMe can do all that and the sample for a few hundred bucks. It’s not like the whole genome is giving you any better information yet, and all that information is going to be public.

Sequencing now is a commodity for scientists researching microbial genomes, but a boutique product for people. In five years, it will be a commodity for people, and you’ll be able to get it from many companies for a few hundred dollars, and one or two for $49.95. Maybe Google Health will do it for free (remember, Google’s founder Sergey Brin is married to a founder of 23andMe, Anne Wojcicki).

So companies will try do differentiate themselves based on the other information they provide—what are you at risk of contracting? What will your kids look like? Where did your ancestors live?

There are obvious difficulties with that strategy. There’s nothing proprietary there — all the good information is public, and the people acquiring the good information don’t work for any of these companies (yet). Plus, very few of the genetic variants have a strong enough effect on health outcomes to be worth communicating. It’s like the wet lab equivalent of Tarot cards.

Reporters that are dipping their toes into the industry are already making note of the basic problem of the fortune teller: Different fortune tellers give you different stories. Companies that do direct-to-consumer genotyping are already facing this problem: the “risk variants” reported by one company’s results may differ from those reported by another company’s results, for the same customer.

I think we can foresee a couple of probable consequences of these problems, as the industry develops. First, companies will need a way to set their offerings apart from the public domain. They will bundle their genotypes or sequence data within a proprietary format so that customers can’t (easily) make comparisons themselves. That will also help to deal with the problem of different “reports” from different companies. If you can’t take your data elsewhere, you won’t be able to get another story.

Second, they will offer up their large datasets to qualified researchers. That wouldn’t be unusual – many researchers consult for industry. With thousands of individuals, there will be the potential to do many kinds of surveys that are now difficult with data acquired publicly. These kinds of results and studies won’t be proprietary, because their results won’t be accepted without peer reviewed publications. But they may set companies apart in terms of mindshare and advance the particular emphases of their offerings. I expect that the company that advances the fastest will be the one that gathers in the coolest stable of young scientists to do association work and population history.

It’s a very good time to get into this field and set yourself apart, because it’s about to go public in a big way.