$1000 genome might be actually really close

[GoodGame]

According to the Wall Street Journal, Life Technologies is promising $1000 genome sequencing to arrive in a very short amount time (possibly under 1 year's time).

http://online.wsj.com/article/SB1000...537379354.html
Alternatively: http://www.pressdisplay.com/pressdisplay/viewer.aspx

Life Technologies Corp., a Carlsbad, Calif., genomics company, plans to introduce Tuesday a machine it says will be able to map an individual's entire genetic makeup for $1,000 by the end of this year. Moreover, the machine and accompanying microchip technology, both developed by the company's Ion Torrent unit, will deliver the information in a day, the company says.

Scientists say that breaking the $1,000 barrier—roughly the price of an MRI test—will accelerate an already fast-moving transformation in genetic discovery and drug development.

This $1000/genome goal has been on-going one as multiple generations of gene sequencing devices have been created. I'm surprised to hear science is so close to realizing it, so soon. I'm slightly skeptical that this is a PR announcement to stir interest and investment, and perhaps inaccurate in the time frame, but it is still provocative news.


Beyond the general benefits (e.g. more effective medications, new medications), this may possibly lead to personalized medicine---possibly bypassing medication all together in favor of individualized genome treatments for disease.


So would you pony up $1000 to have your genome sequenced and on file? (possibly to be of more use to you and your offspring at a later date). Presumably this would involve some painless and harmless donation of your tissue (blood collection and epithelial rubbings.

Would you agree to let that information be used by medical researchers? (A reasonable spin-off is that researchers will get more statistically significant sample sizes to pursue various research projects, and might even pay you to be a test subject, as the cost comes down).

Would you seek out individualized genome therapy for a disease you have?
(an inherited genetic disease, or just as a solution to a more common disease with some genetic proclivity such as Diabetes type 2).

[The_J]

Unless they give some more info, I'll say PR.
Sounds a bit improbable at the moment.

Originally Posted by GoodGame

So would you pony up $1000 to have your genome sequenced and on file?

Wasted money at the moment.
The interpretation is not yet really possible for many things, and when it is possible, they'll going to say "technology is too old, do it again".

That then also answers #3, without interpretation no therapy is possible, and I don't think there'd be already any available.

Originally Posted by GoodGame

Would you agree to let that information be used by medical researchers?

Might be working with human data in the next year, but even while I see the need, I'm going to say no.
I'm not going to pay to let someone else use the data.
Even if they paid it...no. It's probably irrational, I guess.

[GoodGame]

Originally Posted by The_J

Even if they paid it...no. It's probably irrational, I guess.

It's not completely irrational, being that we live in a generation of high technology crime and terrorism. If I were a high-profile wealthy person, I'd probably protect that information as it might be used against me in some way (imagine away for MO's using or targeting genomic information of an individual).

[peter grimes]

I would not, yet.

But I'd be willing to do it anonymously as part of a genomic survey. I'm also fine with genetic screening, since it's a targeted search.

I imagine (hope!) my views will evolve as the privacy protocols become more thorough.

[El_Machinae]

$1000 is certainly my price point. I would happily pay to have my sequence generated and placed onto a disk for me. I think the benefit of getting it earlier is profound, because you can be a customer for any new apps or services that come out. You can pay attention to developments as they happen, instead of waiting for them to happen and then catch up.

Ideally, I'd like an app that flagged articles that report on my sequences. If people did a study comparing the different versions of a gene, the app would tag it for me if one of the tested sequences was mine. Then I could learn tidbits about my genome as the data came in.

I've written a few times that I think that "lay adults" should spend some time learning some genetics, because they're the ones who can get the most immediate benefit. It's a bit like learning how to fiddle around with a computer in the 80s. You'll be on the curve for all the new advancements, and you'll really create value for all your loved ones and friends.

We're not at the $1k place yet, so it means that buying and reading a few books on the topic will create an advantage for those who're willing to learn in their spare time.

[GoodGame]

Originally Posted by El_Machinae

We're not at the $1k place yet, so it means that buying and reading a few books on the topic will create an advantage for those who're willing to learn in their spare time.

Some parts of the industry are claiming that the price point is really really close, as in something you might get for Christmas 2013.
I do think you are on to something with the books. A market for "Dummies guides" for various education levels to potential consumers of a $1k genome will probably spring up on the NY Times book list.

[Boundless]

Well, if we are "really close" who is offering to sequence our genome for $1200?

Or is "really close" still a "closely guarded secret" as to how they propose to do that?

[GoodGame]

Originally Posted by Boundless

Well, if we are "really close" who is offering to sequence our genome for $1200?

Or is "really close" still a "closely guarded secret" as to how they propose to do that?

It's not that close yet Boundless. It's around $5000 per patient for a whole genome sequencing. http://www.utsandiego.com/news/2011/...ng-price-4000/

Small price fluctuations will probably occur after it's here and there's competition.

[caketastydelish]

I think this is great news, but only the rich and/or stupid should pay $1000 for it right now.

Knowing how technology is, when something first comes out, the result isn't exactly beautiful, and it comes at a high price. In 10 years the price may be lower, and the result will be better. 10 years from then, even more so and so on, and so on.

They may make personalized medicine in the future, but I don't see it happening any time soon. Until then, I won't be paying 1000 dollars for this.

[El_Machinae]

I think personalised medicine will be what the 2020s are. By then, the diagnostics and computing will be strong enough, plus personalised compounds.

I guess part of the question is "how much is $1k to you?". If someone is near the median income, I think it would be a good idea to get the information soon, so that you can really keep an interest in new developments, and thrive ahead of the curve. If below the median income, then just remaining educated in the topic would allow pretty good leverage once you can afford the data.

It would be cool to have the data that cost $3 billion to generate a decade ago!

[Bootstoots]

Does anybody have any feel for how cheap whole-genome sequencing can reasonably get? Is there any practical barrier that is likely to slow the better-than-exponential price decline we've seen in DNA sequencing lately? Or should I just sit around for another couple years until I can send them a blood sample with a check for $10? I definitely plan to get my genome sequenced as soon as the price falls within grad-student range.

[GoodGame]

Originally Posted by Bootstoots

Does anybody have any feel for how cheap whole-genome sequencing can reasonably get? Is there any practical barrier that is likely to slow the better-than-exponential price decline we've seen in DNA sequencing lately? Or should I just sit around for another couple years until I can send them a blood sample with a check for $10? I definitely plan to get my genome sequenced as soon as the price falls within grad-student range.

Supposedly the current tech race is beating Moore's law, at least in the past 4 years.

Spoiler:

http://www.genome.gov/sequencingcosts/


I don't think there is any easy answer to that question, especially not in a Moore's law kind of sense. Part of the reason why it's hard to answer is that each sequencing generation represents significant technological advancement in the means of doing things. The generations that follow sometimes reinvent the underlying method of sequencing by using newer discoveries or simply an engineering approach that wasn't previously attempted. For example, the first generation involved just devising chemistry methods for manual determinations. But automating those methods alone doesn't necessarily constitute a second generation. I'd guess that part of the second generation is implementing the most efficient first generation method and modifying it for automation, and so on. But some alternative instruments of a "generation" do go back to the drawing board, implementing more recent biochemistry ideas (e.g. using Polymerase Chain Reaction).

Economic scale will probably play a role in cost too since consumables are a factor in the total cost. Figuring out how the cheapest consumable elements to incorporate into a method, and then figuring out how to scale the manufacturing economy so producing those components is even cheaper, is part of the end consumer cost.
For instance the end might be some sort of disposable sequencer that could do overlapping parts of a project, and fit on microchip device using micro/nano-fluidicsmicro/nano-fluidics into some sort of disposable chip device. Mass produce that with computer-controlled 3D printers and who knows how cheap it'll end up.

$7 a month for unlimited sequencing of choice target genes with limited data streaming to a genome browser?



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

[Narz]

I'd do it for free as part of a study. Wouldn't pay $1,000 though, $1,000 is too valuable.

[MrCynical]

I'd certainly be interested to see what's in my genome - not to the point that I'd pay $1000 for it though. Similarly I wouldn't be willing to pay the costs of sequencing it just for general medical research. I can see that there's some research value in studying genomes in bulk, but not enough that I'd be willing to pay for it personally. I have no particular problem with the information being used for medical research. Keeping it anonymous seems a sensible precaution, but realistically knowing someone's genome is rather useless for any nefarious purpose at present, and I can't see that changing any time soon.

Originally Posted by El Machinae

I think personalised medicine will be what the 2020s are. By then, the diagnostics and computing will be strong enough, plus personalised compounds.

"Personalized medicine" is rapidly turning into a buzzword, used more by marketing than scientists. I'm not saying that it's impossible to do some customization of medicine to the patient, but I've sat through enough clueless presentations on this subject to see how detached from reality most of the predictions about this are. By the end of the 2020s it is plausible that you could sequence the genome of a patient (or maybe a cancer tumour) and determine that of the dozen possible drugs on the market one has an X% better chance of success in a patient with those gene markers.

The main limitation is accurately linking up specific mutations of the DNA with medical conditions/drug susceptibilities etc. Stare at the stats hard enough and you can see some vague trends and tweak the odds a little, but you're not simply going to be able to feed a human genome into a computer and get an accurate diagnosis/predicted response any time soon. There's a reason scientists in this field wince every time they see a newspaper proclaiming "scientists find gene for obesity/religiousness/whatever". We still have nowhere near a complete understanding of what the human genome codes for, never mind the different phenotypes minor changes might produce.

The idea of personalised drug compounds is even more sci-fi. Even if we magically knew the structure of the perfect molecule to treat each patient, it would be impractically expensive and time consuming to devise the syntheses to make all of them. You've only got to look at natural product synthesis to see it can take years to find a way to make a compound even if we know exactly what to aim at. At best, there'll be a handful of possible treatments for a condition on the market, and your genome might give a hint which one has the highest chance of success.

Having my own genome sequenced might give a few pointers to conditions to watch out for. There's certainly a few bits of the family medical history where it would be interesting to know exactly which genes got passed on to me. It's unlikely it would be of much use in treating anything in my lifetime though.

[GoodGame]

Personalized medicine is very believable because it wouldn't be based on natural product synthesis in that it wouldn't be inventing a brand new organic molecule homolog or what not. It's not about making one brand new drug for a rich patient and getting it through FDA trials (or approved on a compassion clause). I agree that if it were based on natural product synthesis at an extraordinary rate, that is sci-fi involving other ideas (e.g. complete knowledge of quantum structure and how to use it for organic molecule synthesis).

The personalized medicine of this thread would be based on genomic science, specifically fine understanding of genomic regulation systems (e.g. variant RNAs like snRNA). It might take the form of an individualized genomic cure, or it might simply be a metabolomics-based diagnostic that is keyed to an individual's genome.

[MrCynical]

Originally Posted by GoodGame

Personalized medicine is very believable because it wouldn't be based on natural product synthesis in that it wouldn't be inventing a brand new organic molecule homolog or what not. It's not about making one brand new drug for a rich patient and getting it through FDA trials (or approved on a compassion clause). I agree that if it were based on natural product synthesis at an extraordinary rate, that is sci-fi involving other ideas (e.g. complete knowledge of quantum structure and how to use it for organic molecule synthesis).

My point about natural product is that even if you did have a full understanding of the human genome and resulting phenotypes, and so knew exactly what you needed, you can't simply go to a chemist and say "the patient needs X mg of this compound by next week". Devising syntheses is a very slow process, and that's without even getting into the whole safety/legal side of things. Best case is that you can cut some corners by having a basic drug that is reasonably effective in general, which can then be customized to some extent by attaching substituents to it with click-chemistry style reactions. That might be feasible in a couple of decades, but you'd still be talking about a handful of variants on each basic drug framework.

The personalized medicine of this thread would be based on genomic science, specifically fine understanding of genomic regulation systems (e.g. variant RNAs like snRNA).

In some ways this is quite a flattering sentiment since this is the area of science I'm working in, but I'm also aware how much hard grind is going into getting an understanding of these systems in general, never mind how they can vary from person to person. We've barely scratched the surface of regulation, RNA splicing, transcription regulation and so on. About the closest we've got in this area is some of the siRNA work, which is still experimental and giving rather mixed results due to non-specific effects. If you're expecting a remotely complete model of how this works by 2030, I'm afraid you're going to be disappointed.

It might take the form of an individualized genomic cure, or it might simply be a metabolomics-based diagnostic that is keyed to an individual's genome.

I'm not sure what you mean by an "individualized genomic cure". As I've said earlier it's simply not going to be practical to come up with an individual drug for every current patient genome, even if we fully understood it. Similarly the idea of reading off a genome, knowing the exact metabolic intake and outake of every cell and hence diagnosing disease is a reasonable line for a sci-fi novel set a few hundred years from now, but I would expect to see it in any of our lifetimes.

[Quintillus]

The WSJ doesn't give much info since it's subscription-only, but according to an MD directly involved in this area who I recently heard, this is legitimately likely (and he doesn't work for that company, either). The figure he gave was $4000/genome right now, with $1000 being likely either this year or next. He also noted that currently it isn't feasible to have everyone's genome mapped even if they had the money, as a matter of scale. But that will change, too.

I'm not sure if I'd do it. I wouldn't today, but when I've saved up a bit of money, maybe. One of the issues I'd have is, what information do I get? If there are things I can do to legitimately increase my health, great. But if I'm predisposed to probably get some disease that we don't really know how to prevent, I'd rather not know that. Ignorance is bliss sometimes.

[Polycrates]

I'm not registering with their website to get the full article, are they talking about whole genome sequencing or exome capture? I would have to assume the latter (which currently sets you back about $7000), which is a bit misleading, although it does cover the majority of genetic alterations that have health implications.

The problem with that is that, as with all of these high-throughput things, any given result it throws up is going to be so unreliable as to be pretty much worthless without independent verification with other techniques (oldschool PCR or whatever). There are a few different "cheap" exome-capture technologies, but all of them apparently manage about a 2/3 to 3/4 coverage of the relevant DNA with enough confidence for it to even be worthwhile as a first pass thing. And they seem to have an awful lot of trouble with GC-rich regions, which excludes things like certain important and frequently-mutated tumour suppressors.
So, sure, it's great as a fishing expedition sort of thing - but the reliability just isn't really there, so it may miss important mutations or throw up alarming false positives, and you can imagine just how much misinterpretation of the results there could be if this became a routine public health thing.
Then of course, how much can you actually say from a particular genetic profile? It may or may not actually mean anything in that particular person's context. There are certain things that are so well-linked that you can take precautionary action (for example, women in families with histories of breast cancer often have precautionary BRCA1 testing and sometimes have preventative mastectomies if the mutation is detected even without ever being diagnosed with cancer), but I can't help but think that targeted sequencing of those particular genes is going to remain a more reliable and useful method for diagnosis for quite a while.

[GoodGame]

Um...I think what WSJ does is make an article free when it's hot off the presses, but once it gets archived they make you subscribe to read it. When I started this thread, the article was completely free, as the article was hot-off-the-press.

I think this is the same article if you care to read the full text: http://www.pressdisplay.com/pressdisplay/viewer.aspx

[Polycrates]

Thanks, but that link's not working for me either
I did a quick google and found some stuff, but I'm still not entirely clear as to whether it's exome or whole genome.
But by sheer coincidental luck, I got an email today that the company is giving us a presentation in a couple of days to try and flog the sequencers, so I might mosey along to that if I remember and see what their spiel is. I'm sceptical, but hopefully they'll at least have good sandwiches or something.