Slashdot Mirror
DNA Assembled Nano-Transistors
Bob Vila's Hammer writes "In an article at New Scientist, researchers at the Technion-Israel Institute of Technology have harnessed DNA to mold a nano-transister constructed of graphite nanotubes coated in silver and gold. The carbon nanotube assembly when completed is a fully working transistor when voltage is applied. The process is ingenious, using proteins from E. Coli bacterium to bind carbon nanotubes to certain sites on strands of DNA. Then graphite nanotubes coated with antibodies connect to the proteins. Finally, silver ions are added to the solution which chemically bond with the DNA site where the protein is attached. Further refinement of the technique is required before full scale production would be efficient, but this could allow the creation of elaborate self-assembling DNA sculptures and circuitry."
So what's next DNA assembled WiFi device inside of our brain effectively using it as a "mobile" storage medium? Probably not only that, but also for doing true multi-task administration in the real world scope.
:)
Just think how quickly one could hack wireless access points around them or a beowolf cluster of brain activity via peer-to-peer. That should rack up some SETI@Home work units completed in no time!
This space is not for rent.
What is the life expectancy of the components? From the article it seems to me (disclaimer: IANAMolecularBiologistOrNanoEngineer) that the organic component is not required after the "wires" are in place but will the DNA auto-repair any damage to the wire?
Couldn't a virus (biological, not computer) be used to re-write the DNA strand that is used to construct the devices, to make different components for sinister purposes?
Is it paranoia if they really are out to get you?
Music is everybody's possession.
It's only publishers who think that people own it.
Fuck Beta
~John Lenno
Further refinement of the technique is required before full scale production would be efficient
/. include some sentence like this. I'm sortof patting myself on the back here when I say this, but hats off to the chemical engineers who actually do the work here. Chemical engineers are an important stepping stone between "oh, cool" and full-scale production, but hardly ever get a mention. In fact, most people have no idea what chemical engineers do, even though you probably scarcely have an item around you that doesn't owe its existence in part to chemical engineering.
It seems like a lot of the "science with potentially awesome applications" posts that get made to
Imagine what the possibilities are here.
Were I in control of this style of circuit manufacture, I would look into creating artifical neurons -- a small CPU core would provide the basic multiply-accumulate-threshold logic on the neuron. Other multiply-accumulate circuits at the synapses or dendrites would provide long-term adaptation functionality needed for learning.
The advantage of a neural net appraoch is that it can work with an inexact network. Standard digital electronics are logically fragile for the most part (i.e., they break if you replace an OR gate with an AND or swap two data lines). Digital electronics depends on highly repeatable manufacturing processes that create exact interconnect topologies. In contrast, neural nets are robust to any-to-any connection topologies and use various long-term adaptation schemes to reinforce or attenuate the connections that are needed.
Thus, you could create a soup of neural node cores, dendrite fragments, axon fragments and synapse units that would self-assemble into a gelatinous brain-mass. Plop the mass on top of a set of electrical interconnects and then train the blob to do what ever you want it to do. Moreover, these nano-fragment brains would be about roughly 10-100 times smaller in each dimension (about a thousand to a million times smaller in volume) than their cellular equivalents.
It could get interesting if we can create human-brain level neural net blobs that fit in a 1 cubic centimeter volume. Neural gel-packs, here we come.
Two wrongs don't make a right, but three lefts do.
They already have joined component built with this method.. but not on the megascale we're used to in modern procesors.
I think circuitry built using this approach would have to be thought about in a fundementally different way.
Fairly obviously (I think) large scale structures like the processors we know and love today would be very dificult to create using this organic approach. A better approach might be to just go for creating very dense, very connected but essentially amorphous 'mats' of computing resource (neuron like units perhaps ??) and treating the whole thing as more like an FPGA than a traditional structured computing device. So the problem becomes not how to grow these things in a particular shape.. but how to persuade the shapeless mass to do something useful.
Would it possible to have these things assembled by protein structures that deliberately mutate at each assembly to provide binding sites that uniquely identify each processing element. That might be a start.
> Different polypetides might make transistors, autonomous clock circuits, chemical-to-electrical battery subunits, wires, tees, etc.
The problem with "biocomputers" is that typical electronic equipment and biological macromolecules have very different properties. Proteins get their "shape" from very specific conditions, including *temperature*.
> An alphabet of tRNA units would carry heavily modified amino-acids and provide both the electrical and structural of properties of the polypeptide.
This is another one of those things that sounds good in theory, but will never work in real life. The reason tRNA have specificity for their *exact* amino acid specificity is because of incredibly precise interactions with the enzyme that links them, and the amino acids. By modifiying any of the three even slightly, you destroy their ability to bind specifically. (in addition, the cost of synthesizing "heavily modified amino acids" could be over $100 per molecule!)
> Each electrical component would have a unique code on each terminal that only binds with the component that it connects to in the circuit. By labelling all the terminii of the components with these specific binding patterns, you the potential for self-assembly.
Biological molecules, unfortunately, don't follow a simple set of unchanging rules. You never know when an already assembled subunit will turn around and bind something that it shouldn't, or when a temperature change will denature a binding site and ruin the whole process.