Digital DNA Circuits

TheSync writes "ScienceNews has a story about digital DNA circuits. The circuits use proteins that activate or deactivate genes on the DNA for control. Since an inverter and an AND gate have been created, any digital logic circuit can now be done in DNA. Moreover, evolution can help make circuit elements work better. There is even a "databook" of BioBricks circuit elements and BioSPICE for biocircuit simulation."

Comments

And, not

[Anonymous Coward]

But does this mean we can store data in DNA using hundreds of bases (latch), instead of a few bases directly?

Oh!

[Quasar1999]

So that's how the neuro-gel packs work in Star Trek... and all this time I thought it was crap!

Seriously though... what's the delay on these things? Comparable to silica versions?

but then...

[Tuxinatorium]

but then, that's not much more compact than a 90-nanometer transistor. Do you know how huge a polymerase protein is?

Re:but then...

[g4dget]

But proteins go into a little 3D bag, while transistors need to be packed near a flat surface with current VLSI technologies.

DNA computing and Cryptography

[jwdb]

Does anyone know of any research into DNA computing and cryptography? I'd expect, given the massively parallel capabilities of DNA, it would be a very useful tool for a brute force attack...

Imagine coding all possible keys as dna, mixing in the message, and pulling out the only possible and logical match -> your decrypt.

Or am I just dreaming?

Jw

Re:DNA computing and Cryptography

[rwiedower]

Even your body doesn't rely upon chemical reactions to accurately predict certain outcomes. Studies have shown that nerve fibers in your arm will often send a "the ball is coming" signal to your brain well in advance of the actual signal reaching your fingers. This sort of predicative function makes complex tasks like walking and talking much easier, but when it catches up to you (like when you fall on the bottom step of a flight of stairs because you forgot how many steps there were) you crash and burn.

The point is, that chemical reactions are very slow. If they were faster, your brain (and your neurons in your arm) wouldn't have to guess. Because they're so slow they'd be very poor at brute force attacks, regardless of the sheer number of cells.

So, yes, you're dreaming.

Illuminate Me

[rwiedower]

Okay, this may seem short-sighted, but if silicon circuits are so much faster, why not simply design silicon-to-carbon interfaces rather than try to redesign the wheel? Unless there's some level of functionality that's not applicable on the silicon side, I don't see why the results of a process couldn't be approximated. In the article, for instance:

It's far easier to describe the schematics of these circuits than to build them for operation inside a cell. For instance, to hook up one gate to the next, the amount of protein produced by the first gate must be the right amount to activate the next gate. And at every step, the output protein must be either very high or very low, to avoid false positives or negatives. It's also necessary to tweak many parameters, such as the strength with which the various proteins and the messenger RNA bind to different parts of the DNA sequence.

If the end result is accomplished simply by having the right protein the right place at the right time, why not build the circuit in silicon and simply train the cell to produce the appropriate protein based on the result of a calculation? Perhaps my ignorance is becoming too glaring...

Re:Illuminate Me

[Mr. Spectre]

Okay, this may seem short-sighted, but if silicon circuits are so much faster, why not simply design silicon-to-carbon interfaces rather than try to redesign the wheel? If the end result is accomplished simply by having the right protein the right place at the right time, why not build the circuit in silicon and simply train the cell to produce the appropriate protein based on the result of a calculation?

They are hacking the instructions written the code that cells already understand. This might allow for programming a cell to produce a particular protein only under certain circumstances determined by the hacked code. How is creating a silicon-to-carbon interface going to get a cell to do what you want it to?

Self-improving circuits

[16977]

The most interesting thing about this announcement is that this guy has been able to use evolution to improve his circuits. I don't expect molecular computers to surpass electronic computers, at least right away -- although they could theoretically perform faster than electronic computers in the short term, any advantage is offset by the time needed to convert the information to human-readable form (by finding and correctly reading the DNA sequence). As the article says, it's better to take advantage of the fact that you can "work with" bacteria. But if DNA computers could repair and upgrade themselves, they would have an advantage that electronics currently does not have. Electronics already is under intense artifical selection, and it can reproduce itself after a fashion, but unlike copper and aluminum, DNA computers can be randomly mutated, and the close homology between computers ensures that some of those mutations will be beneficial.

Re:Self-improving circuits

[rwiedower]

DNA, maybe. DNA is fairly good at reproducing without errors. RNA, on the other hand, isn't that good with errors, but is much quicker. (Ask any virus.)

My thought is this: as soon as the process becomes complex, errors introduced into each cell could produce vastly different results. And the debug process would be tortuous. There'd be no guarantee that a single mutation couldn't bring down the whole system.

Don't know how 'wow'-ing this really is...

[Necromancyr]

Reports of this sort have been coming out for a few years now - basically, all they are doing is a controlled induction of a promoter. It's nothing amazing. Chaining one promoter to express another promoter ad infinitum (or to restrict expression) is already done in nature and used extensively to create transgenic cell lines, bacteria, etc. Hell, they've already developed means to do basic computations with DNA that are more applicable/advanced then this in some respects.

Evolution of Programming Languages?

[jdh-22]

Does this mean that a new bread of modern inovative programming languages will be needed? I am sure that most expirenced programmers would definatily like to do something differently to make the development process more efficient, and faster.

Any suggestions on what you would like to see if a new language was developed for this platform?

proteins

[lazira]

Have there been studies in alternate programming methods/languages for DNA, like there were for quantum computers? DNA logic doesn't need to be sequential- each protein can affect many things at once. It seems rather unwieldy to try to apply conventional logic building blocks, as each gate would require a unique protein and inhibitor- you can't use the same block twice.

Spaghetti Code

[Obiwan Kenobi]

I was listening to NPR the other day which focused on DNA as a computer.

The guy interviewed correllated the DNA genetic map to spaghetti code, a programmers worst nightmare. Apparently all through the genetic make-up of our bodies are "fuction calls" (to put it simply) and pathways that reference other calls and other pathways, over and over upon itself for a hundred million lines.

Its not the listing of the GTAC code (ie, genetic map) that's really necessary. Though of course it plays a part. Its the understanding of such code, what it does and what it controls, where power lies.

The guys interviewed all guessed it would be a hundred years or more before we began truly understanding what "functions" do what in the DNA strand and how it affects the organism in question.

Food for thought.