Hacker News Reader

Comment by lukifer | Hacker News Reader

I was excited about 23andMe, but I decided against it solely for this reason. You can change your passwords, but you can't change your genetics. Whether it's the mundane and real concern of invalidated health insurance claims, or the far-flung hypothetical concerns (future clandestine organizations developing individually targeted assassination viruses), giving away one's genetic code involves too much risk and moral hazard.

(It's a shame, too. I badly want my full ATGC sequence to be a zip file on my hard drive.)

View on HN

Clarification: while 23andMe could fully sequence your genome using the sample you provide (which I assume they bank...), that's not what they actually do currently - your DNA is instead chopped up into short segments and run on an Illumina SNP genotyping microarray (to obtain variant calls at thousands of genomic loci found to be relevant in various association studies).

Your DNA is never actually fully sequenced, because that is currently too cost prohibitive...

> Clarification: while 23andMe could fully sequence your genome using the sample you provide (which I assume they bank...)

I'm not sure they bank it. They offer an upgrade that includes more tests, but to take advantage of it you have to send a new sample. An upgrade that did not require sending a new sample would surely have a much higher conversion rate, so I'd expect they would offer that if it were feasible.

That's interesting. I just assumed that they would, since DNA is pretty stable once extracted and storing nano or micro liters of materials, even at -80C, shouldn't be too expensive considering the upside.

They require a good amount of spit, in order to ensure that they get enough DNA from each sample. It's quite likely that they're not left with much afterwards, and for many of their samples there may be nothing left at all. Banking the DNA is a whole other hassle. Banking the 500,000 samples that they've done so far would require quite a bit of freezer space.

Also, based on their recent statements, it seems that 23andMe doesn't actually do any of the lab work, they contract out to a lab partner. Based on the amount of capital they've raised, I would have assumed that they did everything in house. Anybody know more about this?

"Banking the 500,000 samples that they've done so far would require quite a bit of freezer space."

Maybe one or two -70 freezers full of eppendorf tubes. It's not a significant cost.

I'd never calculated the density of eppendorf tubes before, but it looks like they would fit in 3 cubic meters. Needless to say the freezers I've used have not been nearly that well packed.

My understanding was that the reason they require "a good amount of spit" was actually so that you couldn't steal someone else's saliva and submit it for testing.

I'm 98% sure that when I signed up it was presented as a choice: "Do you want us to bank your DNA for free?". I signed up after they changed SNP arrays (I think it was v2 -> v3 that the above poster is referencing offered upgraded reporting). I'm not sure if the banking was only implemented post v3, or if they had some other reason to re-collect spit despite having some DNA banked.

Your DNA is never actually fully sequenced, because that is currently too cost prohibitive

Q> Does anyone know the actual/current pricepoint for this?

going by wiki: https://en.wikipedia.org/wiki/Whole_genome_sequencing#2012

1000$ - 5500$

http://en.wikipedia.org/wiki/Full_genome_sequencing

The page cites only 69 people so far had it fully sequenced, and it cost about $100k a few years ago

That figure is only for publicly available sequences. The cost is now under $5k, and I'm sure thousands, if not tens of thousands, have had it done. (Including me!)

Currently about $3,000 and falling.

See http://www.genome.gov/images/content/cost_per_genome_apr.jpg

Not quite... that it the raw sequencing cost and doesn't cover the prep work required, or likely the computational resources required for processing. It is actually almost to the point where the computational costs are more than the costs of raw sequencing.

And, for the places where you can get a $3K genome sequenced, it isn't likely that they will service outside (non-research) customers.

Custom-made virus would probably cost millions for an uncertain result (delivering a virus into your bloodstream would require some work). A drug-addicted thug with a lead pipe would probably not even get the cost into four figures. I think your priorities are misplaced. Unless you are protected on a presidential level, if somebody decides to assassinate you the simplest plan would probably be the most effective one.

I've read too many cyber-punk novels, I know. :)

I'm more thinking along the lines of what may or may not be possible in 30-40 years, and even then it's an admittedly extreme example.

I'd really like to see some better analysis of that. The ACA has made the first point mostly moot in the US (as it already is in most of the developed world), and... let's just say that I'm willing to take my chances with the "assassination virus" thing.

Seriously, what's the real danger here? I just don't see much. I paid 23AndMe and was generally pretty amused (and sadly unsurprised; not much of anything interesting turned up in my profile).

My examples are endpoints on a spectrum. The core point is that while risks are low now, giving up one's genetic data is a one-way decision that can never be undone. It's a gamble, considering the unknown unknowns of both technology and future public policy.

I suppose in the case of future clandestine organizations developing individually targeted assassination viruses, they would be able to obtain their own sample from the individual.

What I don't understand is: People use 1Password etc to generate arbitrarily secure passwords for Facebook but agree on sending their DNA password (even though its 1B chars) on a postcard to a company (with unknown security credentials and intentions) just to know that they are 29% east-european and probably sleep better at night. Sigh...

Unless you go to extraordinary effort to keep your DNA confined, you almost certainly leave samples in publicly accessible places hundreds of times a year.

There's a difference between leaving a door unlocked vs. mailing a stranger the key.

How long do those samples last?

DNA lasts a long time. Just watch any one hour crime drama. :)

Actually, it could last an arbitrarily long time if it wasn't exposed to UV or cleaned up. The hard part would be distinguishing your random DNA from that of everyone else.

It's not a question to keep my DNA confined, but rather making it random. If somebody wants to track my DNA, yes they can do that - here in Switzerland, far away - but I wont send it to them on a silver plate.

Reply on news.ycombinator.com