IBM's ultra-cheap DNA Transistor dream could lead to personalized medicines, confusion
By Darren Murph 
posted Oct 8th 2009 9:45AM
posted Oct 8th 2009 9:45AM
Are you ready to get your nerd on? No, seriously -- are your rimmed glasses and pressed slacks at the ready? Good. IBM has just announced a full-on research project that intends to drive the cost of DNA sequencing down from millions of dollars to under $1,000. The reason? An ultra-cheap, silicon-based DNA Transistor could essentially "pave the way to read human DNA easily and quickly, generating advancements in health condition diagnosis and treatment." Moreover, it could eventually lead to personalized genome analysis and personalized medicines, meaning that your weekly dose of pills may literally have your name written on them. Just think -- with this breakthrough in place, you might just live long enough to see the Robot Apocalypse. Fun! Video's after the break.[Via NY Times, thanks Serge]
WOW! That would be really cool.... I wonder how much my co-pay, deductible, co-insurance, premium and out-of-pocket would be for this one. This technology is really cool. I think nano-medicine-droid is uber.
Not trying to open a can of worms here.
Gattaca.
Sign me up I want silver eyes like Pitch Black with the super healing powers of Wolverine...
=P
Cool tech but will mostly fall in the wrong hands will. If you know anything about modern medicine you would also know it's one big scam.
The only bad thing I can really think of about this would be being able to know that you have, say, Huntington's Disease, long before you ever show symptoms and living your life slowly awaiting the onset of your disease or wondering whether you dropped something because you were being clumsy or because you were beginning to get symptoms of the disease. Do you see what I'm saying? Although recent laws have made genetic discrimination illegal, you still have to wonder whether your insurance company will still cover you as they did previously when they know that you will get some debilitating disease in your lifetime. Things to think about.
nice
Just give me the red pill already. You can take that blue pill and shove it.
Actually, the red pill is a suppository.
So either way there'll be shoving involved.
I guess IBM will be racing professor Quake from standford.
About a month ago, he published a paper on a process he used to sequence his own DNA in less than a week, for under 50 K. Dr. Quake is begining to commericalize his technology and hopes to also break the sub-1000 dollar barrier. I think there are about 3 other research labs try to commericalize technology for similar purposes.
I know a lot of jokes will be made about this on Engadget, but the potential applications of these are massive. Like world changing, Nobel prize winning massive. Already some physicians are using personalized medicine to treat AIDS significantly more effectively then before. This type of medicine will be able to screen for cancer and other inherited diseases tremendously fast. A number of us in the scientific community are trying to look for biomarkers that are a result of DNA mutations, but going to first principles has to potential to be faster, cheaper and more effective.
Sweet. I never smoked before cause I didn't want to get cancer, but now I can smoke, look cool, and then just take a magic DNA pill in 10-20 years and be cured ^_^
THANK YOU SCIENCE!
This is not a lot more than an IBM commercial. Solexa and ABI (and other) next generation sequencers are on the market and revolutionizing biological sciences. IBM's idea sounds cool but also seems like the proverbial hammer seeing everything as a nail. They make semi-conductors. What would be cooler is if they used DNA molecules as nanotech transistors to control electrical paths (that is a transistor made of DNA rather than a DNA transistor). Maybe DNA sequencing is a field that Nvidia should get into (assuming Intel isn't already there)?
So now I can find out approximately when and how I will die? *Barring anything like getting in a fatal accident
what a great comment robpetrin!
Thanks!
The countdown to Joy has ended.
Of all the diseases, cancer is the one that will require the most personalization of treatment (especially when it's late stage).
What we need is the ability to cheaply sequence the genome of individual cancer cells.
Cancer drugs will need to be designed on the fly in the future. As in, if you have a certain cancer, the drug may have to be designed specifically for your cancer based on its genetic profile. What I mean is that they will have to take a random sample of your cancer cells, sequence them INDIVIDUALLY .. then figure out what drugs to give you in combination so that all the cancer cells can be destroyed with a mathematically infeasible chance of resistant mutant survival. In some cases this could mean developing monoclonal antibodies or small molecules tailored only to an individual.
There will need to be a mechanism in place for allowing this treatment (require virtual toxicology profiling? Require some animal test? Require the dosage to bbe given in very small amounts first (dangerous & could result in sequential resistance mutations .. that will defeat the combination therapy and make the cancer highly resistant).
If you target 4 pathways, a treatment resistant cell would need to have mutated versions of all 4 pathways. This is extremely unlikely, because like you said even under current modalities treatment resistant cells are unlikely.
Imagine you are targeting 4 pathways: A, B, C, and D. Since a cell is very complex, there's actually tens of crucial targets to choose from.
Let's say there is a 99% probability that pathway A is mutated in such a manner that your drug is ineffective .. Those 1% of cells would need to ALSO be simultaneously resistant to pathway B, C, and D drugs. Assuming those pathways are unmutated in 99% of cells the chance of a cell that is simultaneously resistant to all 4 is .000001% (1/100 times 1/100 times 1/100). One in hundred million cells .. which admittedly is not enough .. but then if you target the mutated pathways too (8 targets total) .. then you're talking big impact of 1000 trillion cells.
Can some cancer cells survive? It's mathematically infeasible .. but maybe .. but definitely not in most cases . so most (99%) people will get cured completely the first round. Those few who had a few cancer cells survive will need a second round.
If the combination therapy method is not used what happens is .. pathway A is targeted .. then 1% of cells remaining are resistant .. they grow and multiply .. when they mutiply the chance of a mutant pathway B's existence increases. So then when you treat pathway B .. there is a good chance of mutants. then the same thing happens for C followed by D. It's virtually guaranteeing failure.
Usually the probabilities are a fraction of a percent. A triple combo is usually enough.
An example of effective combo therapy: the AIDS cocktail is typically only 3 or 4 drugs .. and each of those drugs are not very effective given alone .. but given simultaneously it is enough to wipe out the HIV to undetectable levels (the drugs can't hit the provirus or viral reservoir unfortunately, so it can't wipe it out).
One problem is that each cancer is heterogeneous. You'd have to sequence every single cell to cover all of the possible variants. Another problem is that we don't have too many drugs that are selective for pathways. Block the PI3K pathway and you kill normal cells. Block the Erk pathway and you kill normal cells, etc. Lastly, cancers constantly generate new mutations (often the initial hits are in genes that scan and repair DNA damage) hence tumor cells are very good at working around drugs (unfortunately). Personalized medicine will have an impact but cancer will be a lot tougher nut to crack than most other diseases.
It seems odd that a venture capital style pitch with such little proof of concept is getting the NYT and Engadget attention. I guess IBM's involvement is noteworthy, but there are many of high-throughput low-cost DNA sequencing projects in the works and some are actually close to production. The Gates Foundation and other big tech players are even sometimes involved. This project is in the earliest concept stage. I would expect some Engadget inquiry into how this is even a transistor. The concept that DNA is semi-conductive is highly controversial and frankly unlikely. How does this sort of project get no scrutiny and such high profile exposure?
READ THE ARTICLE. Engadget left out the good stuff
Just wait until insurance companies require you to hand over your DNA, for "personalized" coverage...
That's his DNA sequence, which incidentally, explains everything.
They exist, and it's called FPGA's.
The comma generally comes after the word "which," not before. They teach that in 10th grade English which, incidentally, you must have failed.
Also, my DNA sequence has way more M's than that.
Sequencing DNA genomes has already been cut from $ millions to $ thousands. The technological advances over the past 5 years have been stunning (about 1000X faster/cheaper than Moore's Law). It cost $500 million to sequence the first human genome 10 years ago. It now costs about $5000 (although coverage varies). Not that we know what much of the data means..... There's only 20,000 human genes and two thirds of these are guesswork in terms of what they do.