[ymc]

I'm not convinced about that. I see prices of about $98 per sample for the HumanOmniExpress kit, and that's without any "friend" discounts [can't get a price for Express+]. I would expect that 23andMe would have a frequent flyer discount for Illumina, and most of the sample prep should be robot-driven at the scale I'd expect them to be working at.

Based on a single discussion I heard at a small ethics conference a few years back, I'd expect that 23andMe will keep on trucking on and chasing the money, regardless of what the FDA says. Sure, they'll work with FDA as much as possible (or as much as necessary), but it's not likely to stop them from doing the tests.

If they can't market directly to US consumers, they'll find some other way to get that testing done (see Tesla in Texas), or just skip the US market entirely if it stops being profitable (unlikely).

But they are not doing the experiment themselves. They sent samples to a CLIA lab to do it, so there is an overhead there too. Anyway, I think we can agree that their margin is really thin at best.

I am curious what will happen to their web site. Is it a marketing tool? Will it be shutdown? Or if you are access it from an US site, you will be directed to a blank page with a login to access your exisiting data?

[rskr]

Maybe they will make the web site open source.

[Genohub]

A few good reads:

Why the FDA can't be flexible by 23andMe, by law

23andStupid: Is 23andMe self destructing?


- Genohub

[Genohub]

Excellent New Yorker piece: THE F.D.A. VS. PERSONAL GENETIC TESTING

[Genohub]

23andMe Stops Offering Genetic Tests Related to Health

23andMe Stops Offering Genetic Tests Related to Health

[gringer]

So, they're still doing the same things internally, still selling the kits, still letting people download them, still letting people see their association results if they purchased prior to the FDA anger, and still letting people do the ancestry stuff. The only thing that has changed is that 23AndMe will no longer provide a web service for new tests that includes medical association analyses and disease prediction.

I envisage a bunch of start-up "analyse your 23AndMe data for less than the cost of a doctor's visit" websites (assuming they don't already exist) in the near future, shifting from a single website with a consistent analysis policy that is easy to regulate and standardise to multiple websites from different companies that all have different ideas about the most appropriate statistics for calculating probabilities.

[Genohub]

Promethease is an example of one of those services: https://promethease.com/ondemand

23andMe would have to convince users to go offsite for health analysis. Not sure if the general public would have interest anymore.

- Genohub

[ymc]

23andMe Stops Offering Genetic Tests Related to Health

Interesting. As I predicted in earlier post, they take out the health-related stuff for now. To my surprise, traits related stuff is also removed as well. I wonder how they can grow their business in the near term.

A sad day for consumers indeed.

[ymc]

So, they're still doing the same things internally, still selling the kits, still letting people download them, still letting people see their association results if they purchased prior to the FDA anger, and still letting people do the ancestry stuff. The only thing that has changed is that 23AndMe will no longer provide a web service for new tests that includes medical association analyses and disease prediction.

I envisage a bunch of start-up "analyse your 23AndMe data for less than the cost of a doctor's visit" websites (assuming they don't already exist) in the near future, shifting from a single website with a consistent analysis policy that is easy to regulate and standardise to multiple websites from different companies that all have different ideas about the most appropriate statistics for calculating probabilities.

Well, I don't think anyone US-based will now dare to run such start-ups now.

[ymc]

Promethease is an example of one of those services: https://promethease.com/ondemand

23andMe would have to convince users to go offsite for health analysis. Not sure if the general public would have interest anymore.

- Genohub

Looks like they are Maryland based. I suppose FDA should also go after them for fairness' sake?

[dpryan]

Looks like they are Maryland based. I suppose FDA should also go after them for fairness' sake?

If they haven't done the needed validation and aren't presenting things in a meaningful way then I'd be happy to see them shuttered.

[ymc]

Does 23andme have a competitor that provides a service that involves a physician? If there exists such a service, does that make that service a "Lab Developed Test" that is acceptable to FDA for now?

As a matter of fact, OncoType DX and Foundation Medicine are both LDTs.

[Richard Finney]

I guess the FDA will probably be pulling pregnancy tests next as "patients" may make irrational decisions based on the information without consulting a physician.

As suggested, a 3rd party could provide the search of NIH's gwascatalog.txt
No big. It's pretty easy to put some CGI/CSS/HTML5 + stock images on the following query ...

cat ~/gwascatalog.txt | cut -f8,15,21,22,28 | sort -k3gr | grep rs2075650
AB1-42 TOMM40 rs2075650-G rs2075650 3E-7
Age-related macular degeneration TOMM40 rs2075650-A rs2075650 1E-6
Age-related macular degeneration TOMM40 rs2075650-A rs2075650 8E-8
Alzheimer's disease biomarkers TOMM40 rs2075650-? rs2075650 1E-6
Alzheimer's disease (late onset) TOMM40 rs2075650-G rs2075650 5E-36
Alzheimer's disease TOMM40 rs2075650-G rs2075650 1E-295
Alzheimer's disease TOMM40 rs2075650-? rs2075650 2E-157
Alzheimer's disease TOMM40 rs2075650-? rs2075650 2E-16
Alzheimer's disease TOMM40 rs2075650-? rs2075650 3E-11
Brain imaging - rs2075650-? rs2075650 NS
Cardiovascular disease risk factors TOMM40 rs2075650-G rs2075650 2E-14
Cardiovascular disease risk factors TOMM40 rs2075650-G rs2075650 4E-8
Cholesterol, total TOMM40 rs2075650-G rs2075650 3E-19
Cognitive decline TOMM40 rs2075650-G rs2075650 2E-8
C-reactive protein TOMM40 rs2075650-? rs2075650 1E-7
Longevity TOMM40 rs2075650-? rs2075650 3E-17

[dpryan]

Right, because there's no difference between a test that advertises its percentage reliability on TV commercials and one for which the customer actually needs a relatively sophisticated understanding of statistics (multiple testing, relative risk, etc.) to not misinterpret.

[Richard Finney]

Does a customer need a relatively sophisticated understanding of the internal combustion engine to buy a car? Of the odds that he will die in crash? No.

Does the purchaser of any good or service require sophisticated knowledge of costs and benefits?

Can't someone who has the BRCA1 mutation decide "well, I need to start mammograms at age 30", or if they don't "well, I'll wait until later in life"?

More information is good. People taking personal charge of their health is a good thing.

Even if there's strong evidence that knowledge is bad, is it right to outlaw it?