This is apparently a thing that actually exists, not some Drexler fantasy.Using the DNA origami method, in which complex three-dimensional shapes and objects are constructed by folding strands of DNA, Douglas and Bachelet created a nanosized robot in the form of an open barrel whose two halves are connected by a hinge. The DNA barrel, which acts as a container, is held shut by special DNA latches that can recognize and seek out combinations of cell-surface proteins, including disease markers. When the latches find their targets, they reconfigure, causing the two halves of the barrel to swing open and expose its contents, or payload. The container can hold various types of payloads, including specific molecules with encoded instructions that can interact with specific cell surface signaling receptors.
Douglas and Bachelet used this system to deliver instructions, which were encoded in antibody fragments, to two different types of cancer cells -- leukemia and lymphoma. In each case, the message to the cell was to activate its "suicide switch" -- a standard feature that allows aging or abnormal cells to be eliminated. And since leukemia and lymphoma cells speak different languages, the messages were written in different antibody combinations.
"We can finally integrate sensing and logical computing functions via complex, yet predictable, nanostructures -- some of the first hybrids of structural DNA, antibodies, aptamers and metal atomic clusters -- aimed at useful, very specific targeting of human cancers and T-cells," said George Church, Ph.D., a Wyss core faculty member and Professor of Genetics at Harvard Medical School, who is Principal Investigator on the project.
All those sci-fi stories where someone makes a device that only kills the enemy species/race/tribe. We have one of those now.
I see no way this could possibly be abused.
Not really. This thing works, but it doesn't reproduce, so it still makes more sense to give existing leukocytes its capabilities, which is very much the same thing as an engineered vaccine. There is where you're going to see some real amazing cancer therapies coming up in the next 3-5 years. And a few years after that probably some in vitro heart and vascular tissue injectable stem cell therapies which could start to cut in to the other big three causes of death.
Including stroke?
To save others from immense effort with teh googles:
WHO Fact sheet N°310 : The top 10 causes of death : Updated June 2011
1. Ischaemic heart disease: 12.8%
2. Stroke and other cerebrovascular disease: 10.8%
3. Lower respiratory infections: 6.1%
...
4. Chronic obstructive pulmonary disease: 5.8%
5. Diarrhoeal diseases: 4.3%
6. HIV/AIDS: 3.1%
7. Trachea, bronchus, lung cancers: 2.4%
8. Tuberculosis: 2.4%
9. Diabetes mellitus: 2.2%
10. Road traffic accidents: 2.1%
Obviously he meant 3 big causes of death among first world people who could actually afford such treatments.
"Stoke" is also number two, even ignoring the little swarthy paupers we'd all rather not think about. In fact, it's only number 6, at under 5% of the low-income countries. So the more you ignore them, the strokier the world gets. As you could see right in the linked document.
Regardless, I'm ready to be surprised about how this neato delivery system could treat the (large) class of brain clotty fatal awfuls, but it would be a surprise.
After reading this article, I wrote a lengthy response, based on a cascading series of horrible realizations.
https://plus.google.com/117565422980496432687/posts/eZCMzDP3Tku
The thing I don't understand about molecular devices like this is, how do they make them? These are made of DNA, but what does that DNA code for? I gather they're not self-reproducing (avoiding the Von Neumann Gray Goo Nightmare) so does that mean that there's some "glider gun" pumping these things out, and if so, how did they make that? Or do they just shake a bag of spit and blood an infinitely long time until the properly-shaped machine falls out?
Is there a GCC toolchain that eventually gives you DNA? Or is it more like Lego Mindstorms?
GCC Toolchain :-) That's great.
"Like I've got the time to build and maintain some damn monkey-patched cross compiler - call me when me back when you've got it self hosting. And install dropbear for me this time - I'm not running a tty over an atomic force microscope again"
Ahh, someone needs to sciencefictionize this...
More like Lego Mindstorms, except you have to calculate your pieces out in advance. There are two main branches of synthetic biology today: biological circuit design, which is really just fancy traditional biochemistry, and "scaffolding." Take a close look at the figures in this paper on scaffolding. Especially the robot waving at you. After they got it figured out on a computer, the actual lab work involved was: (a) upload the sequences to the DNA sequencer; (b) mix the resulting sequences in a test tube; (c) shake well; (d) wait a few minutes; (e) examine the resulting goop under a microscope.
They just use a solid phase synthesis machine (plus some tricks) to make an arbitrary DNA sequence. Once you've made any sequence you can amplify it if you want using the polymerase chain reaction. You can literally email an arbitrary sequence that codes for nothing at all to any one of a bunch of companies and they'll ship you back a bunch of DNA in a vial a couple of days later. There are limits on how long the sequence can be because of technical issues, but that length is getting much longer every year.
The exact mechanism by which solid phase synthesis works gets a little technical but you can find an explanation in most good organic chemistry textbooks. Basically you have a bunch of tiny polystyrene beads and you prime them with the first DNA base with a "protecting group" at one side. Then you bathe the polystyrene beads in something that removes the protecting group, wash that out, and then add the next base + protecting group. Now you have two bases in a row with a bead at one end and the protecting group at the other. Then you repeat, removing protecting groups and adding a new base, one after another. The polystyrene beads are there literally to anchor the DNA as it gets made so that it doesn't get washed away -- otherwise the yield on the reaction would be so low that you couldn't get hundreds of base pairs in a row. The whole process is automated these days -- a computer runs it.
As for how they figure out what sort of sequence they want in order to make the DNA fold to produce the sort of DNA micromachine they want, they effectively use a compiler for that -- the technique was developed by a couple of guys at Cal Tech. The technique is literally called "DNA Origami".
I can't help wondering about the security implications of just sequencing whatever someone sends you in the mail. What if it's a prion disease? What if it causes "MAKE MONEY FAST" to appear in Petri dishes?
...the sort of DNA micromachine they want...
Great, now I have the image of John Moschitta rattling off DNA sequences in my head.
Just like we have human-readable languages that can be translated into executable code on processors, I can easily imagine a time when some human-readable, textual language with deep semantics will be available to DNA-based hardware builders. In some ways, I hope to live to see that become widely available; on the other hand, see above.
Oh, and FYI, Drexler's stuff isn't "fantasy", it is just decades out. We don't have the tools to build the molecular machines you find in "Nanosystems", but if we could build them they would work. The problem is, we don't yet know how to build them, but we're getting closer and closer. It will probably happen this century, though it probably will not happen soon.
I think if there's anything the last few decades of technological change have taught us, it's that "a few decades out" is so hard to predict with anything approaching accuracy that it might as well be Edgar Rice Burroughs writing about Mars.
You mean you're not enjoying your wall-sized OLED monitor?
You'll be able to get those wall-sized this year, provided you live in a capsule hotel. In two years this comment will look like a total non-sequiter.
Has anybody worked out the control issues with Drexler-style nanotechnology? Once you've assembled your atoms, how do you control what it does?
No. There is no control system more complex than a handful of finite state machine states at that scale which is even vaguely allowed by the laws of thermodynamics. If there was, life would have found it a long time ago.
I'm glad you said that, DFB. That tells me that human innovation will only be able to build on what is already widely used.
Actually, life did find it, its called RNA. Drexlerian style machines at the smallest scale would be programmed with "tapes" similar to messenger RNA. Larger machines would control the smallest of them, still larger computers/machines control these, etc. Nanasystems is quite complete as a design doc for nanomachines.
Life does it all the time. A grey whale is built from the ground up using nanomachines, from molecular scale instructions (DNA).
Last time I looked at Drexler, he was contemplating a considerably finer grain.
Hey, awesome, the organic fair-trade handwoven version of "nanobees". (1. Was that a Previously? 2. It was annoying trying to find an actual non-pay article about those mentioning they found a way to get them to respond to surface proteins or such, even though that was front and center in the Nova episode where everyone heard about them. 3. Why the fuck is Firefox 10's spell-checker reliably believing 'center' is not a word?)
Meanwhile, some already-human-approved chemo drug maybe scrubs out Alzheimer's plaques. This grim meathook future of ours is getting kind of badass.
this shit is serious nobel material