My first job out of college was with a startup that did health risk assessments. Part of this involved the use of new genetic markers for disease. This was decades ago (yes before the Internet even!) so it was early days for genetic testing. We had an expert system that put a questionnaire in front of customers and charged them for the resulting recommendations and any tests we did.
Two 'hot' technologies of the time: expert systems and genetic testing.
We learned:
1. people want >answers< not statistical results.
2. people told they have a significant probability of a disease get unhappy. Do that enough times someone gets unhappy enough to sue. So your follow-up debriefings have to be done carefully and in person - very expensive.
The lawyers that ran the company did not see the cost of lawsuits, on top of the high cost and error rates for genetic testing at the time, as a deal breaker going in.
Good to see 23andMe make what looks like a better attempt at this market.
Some great conversation on Louis Gray's Buzz, including a comment by the 23andMe co-founder:
"This is one of the most thorough and accurate summaries I've seen (I'm the co-founder of 23andMe). One small point...the testing technology 23andMe uses is highly accurate, it's just the interpretation of what the datapoints means which is, in some cases, early and subject to revision (this would be equally true if you sequenced the entire genome vs. genotyping, which is the SNP-based technique used by 23andMe)."
It's not - it was an option during sign up. It's just useful in case they have new features in the future and need to reanalyze your sample you won't need to send in a saliva sample again.
We got the family package and were pleased with the results. My wife was even offered a free package to give to relatives because her genome was so unique to them.
Note that you can download the raw data from 23andMe and go to the public http://www.snpedia.com and use the app they have (Promethease) to get lots of additional details that match your individual SNPs/genes.
Most of the SNPedia data (that 23andMe didn't have) was from newer studies without massive amounts of confidence, but it was still fairly interesting.
I was surprised to see my above comment copied to the parent blog link as if it was a comment on the blog itself. What's the software driving that? Very disconcerting and pretty cool at the same time. Makes me hesitate to comment on YC, though, when the comment could be taken out of context when read elsewhere.
I'm using a plugin by YC company BackType. it's called BackType connect. I agree that there are some issues with context, especially when it doesnt currently thread stuff. Which is why I opted for those comments to be pushed to the bottom, after regular comments on my blog.
Actually thinking about it, it should be possible to purchase it with an anonymous prepaid card and use a fake name? anyone knows good anonymous credit cards that I could get in Asia or Europe?
Very cool. I wonder if they continuously update the correlative research and its implications for existing customers. That seems like the portion which needs the most work and I would hope you wouldn't have to pay $500 every few years to see the latest results.
I would definitely subscribe to an annual service that would ping me every time a new report came out with implications for me.
The statistics and pieces of research get updated regularly, but not often. More importantly, you can download your SNPs and go to SNPedia and look up whatever you want by yourself.
Very interesting stuff. Probably the greatest value will be the wealth of data gathered from the biobanking you agreed to. Glad you consented to that. DNA is mind-boggling, and Watson & Crick made the most import discovery of the Century elucidating its double-helix architecture in 1953, in my opinion.
Two 'hot' technologies of the time: expert systems and genetic testing.
We learned:
1. people want >answers< not statistical results.
2. people told they have a significant probability of a disease get unhappy. Do that enough times someone gets unhappy enough to sue. So your follow-up debriefings have to be done carefully and in person - very expensive.
The lawyers that ran the company did not see the cost of lawsuits, on top of the high cost and error rates for genetic testing at the time, as a deal breaker going in.
Good to see 23andMe make what looks like a better attempt at this market.