Slashdot Mirror
Researchers Build Logic Gates With RNA
Ars Technica reports on research out of Cal Tech where scientists were able to create logic gates out of RNA molecules. Thus far, they've demonstrated AND gates and OR gates, with work proceeding on more complicated systems. The work shows promise for ability to easily detect the presence of particular chemicals. The abstract from the scientists' paper is available at Science. Quoting Ars:
"Detecting tetracycline isn't especially interesting, but RNA that binds to specific small molecules is actually relatively easy to make; repeated rounds of amplification and selection for binding can evolve these RNAs in a couple of days. This means that, in a matter of days, researchers can grow yeast colonies that glow in response to a variety of chemicals, or even to combinations of chemicals. More complicated circuits should be possible if the ribozymes are inserted into messenger RNAs that encode transcription factors, which could, in turn, regulate genes that encode yet other ribozymes."
But how long before it runs Linux?
Completely wrong track I'm afraid.
Skynet is more cell phones, cash registers and anything with a system.
Matrix was more humans are batteries, but in this case we get to be both networked pods and large D cells.
Yes, it could be said that the matrix villains are far more green then the terminators. Marketing research shows that hippies prefer matrix type overlords.
"You should always go to other people's funerals; otherwise, they won't come to yours." -- Yogi Berra
Perhaps the results of this research can be used to create biological instances of the satisfiability problem.
If satisfiability can be reduced to DNA transcription in polynomial time, then we could genetically engineer colonies of randomly poisoned cats in boxes to solve NP-complete problems.
There are no karma whores, only moderation johns
Why are they working on more complex systems? If they already have AND and OR, all they need is a NOT and they can make any other type of gate
1. Actually, I don't know if they really need a cell.
Even from the viewpoint of life evolution on Earth, it all started with some self-replicating ribosyme that "lived" perfectly well in the soup of aminoacids and nucleotides around it. The cell was just an increasingly complex test tube around that reaction, complete with increasingly complex ways of regulating the exact composition of the contained drop of sea water.
I can see how that was an advantage to evolve, in a primordial soup that was hit and miss anyway and probably (very slowly) degrading in quality over time. But in a lab, we can do that regulating artifficially. Admittedly, using a cell might be cheaper, but we can do without it too.
And indeed there is plenty of organic stuff we already do without a cell. E.g., detecting certain DNA sequences is done via enzymes which bind exactly to one sequence, and start replicating it until it's enough to be detected. We don't really build specialized cells for that.
2. Actually, to me another aspect is more interesting there: the fact that it's all done with RNA.
Proteins already _do_ exactly what these guys seem to do: bind only to certain mollecule configurations, but not to others. You can see it as logic operations and whatnot, but really it's all chemistry and that's all it does: bind only to certain mollecules, but not to others. It's a bit like saying that a keyring with two keys is a mechanical OR gate: it unlocks a lock that matches either key 1 or key 2. It's simultaneously technically true, and a bit misleading.
But there's a more interesting aspect to it: your body usually uses proteins for that, and DNA/RNA is just a way to encode a protein which will actually do the matching. E.g., those enzymes I mentioned, are proteins. They do all the heavy duty chemistry, from processing the cell's "food", to regulating what goes in or out, to destroying all chemicals which are non-polar and pass right through the cell wall instead of being regulated by the protein valves on the wall, to movement, to DNA repairs, to regulating what other proteins are built and where do they go.
As long as that's all the model we know, that needs a rather complex initial configuration for the start of life. You need something that's capable not only of replicating itself, but also of encoding proteins. It's already a bit too big an incredible machine, and appearing out of nowhere, even after billions of years and trillions of tries per second, still is a damn improbable event.
But that everything can be done via RNA only, that opens a whole new possibility. We already know that RNA can replicate itself. If it can also take the functions of a protein, offers a much simpler initial configuration for life. It's entirely possible that assembling proteins came later, as a better replacement, much like DNA later replaced RNA as the encoding of choice. The first cells could have been RNA-only, but could still have a metabolism and be able to regulate themselves well enough.
I find that fascinating.
A polar bear is a cartesian bear after a coordinate transform.
You didn't really get it.
With a NAND (or a NOR), you can implement any logic function.
NOT(x) = NAND(x,x)
Now that we have a NOT gate using NAND only, we can use it to implement AND:
AND(x,y) = NOT(NAND(x,y))
Using AND and NOT (which only use NAND), you can implement a OR:
OR(x,y) = NOT(AND(NOT(x),NOT(y)))
You can't do this with AND, OR, or combinations of the two. Specifically, you can't use them to implement a NOT.
Why restrict them to discovering a nand implementation? I was going to ask how close they were to implementing "not".
I'd rather ask the less restrictive question of how close they are to implementing a functionally complete set of gates in their process technology.
So for example this could be any of {nand}, {nor},{and,or,not}.
After all, it could be that in the RNA domain, building things out of all NANDs just isn't as efficient (in whatever sense they mean) as in static CMOS.
Yes, yes, it looks and sounds cool and that's precisely why it was accepted into Science, but if you look closer at the results then you'll be greatly disappointed.
Why? Their RNA switches don't really perform as logic gates. When we think of logic gates, we think that the "signal" -- voltage in the case of electronic circuits or the production of a reporter gene in this case -- will have a clear difference between the "ON" and "OFF" states... 1 and 0. Electronic circuits are designed so that the voltages that represent the 1 and 0 are very, very different.
Unfortunately, these RNA switches do not have a clear separation between the ON and OFF states. The authors manipulate their data using disingenuous techniques that mislead the reader (to say the least!). Let me give you an example. When the RNA switch is ON, then the gene expression reporter will have an output of (for example) 1030. When the RNA switch is OFF, then the reporter will output 1000. The authors will report this as a 30 "unit fold change in device" or some other crappy made up unit.
Unfortunately, you can't use this RNA switch to DO ANYTHING that the authors say it can do. If you put another gene under control of this RNA switch then the "OFF" value will be so high that the gene will effectively be "ON". When the RNA switch is turned "ON" the change in gene expression will be so small compared to the baseline that the actual physiological effect will be negligible.
This is not the first time that the authors have mislead readers by manipulating their data (see their previous PNAS paper) and the LEADERS in the RNA switch field have vigorously complained that their work does not actually solve the RNA switch problem --- it just changes the way that the data is analyzed to make it appear that some problem was solved. Eventually, other scientists will discover the falsehoods and heads will roll.
This is a sad day for Science (both the journal and the pursuit thereof).
This means that, in a matter of days, researchers can grow yeast colonies that glow in response to a variety of chemicals, or even to combinations of chemical
As a homebrewer, there are lots of chemicals that show up in beer, some good, some bad. It would be great to modify a strain of yeast that would glow when diacetyl or some other chemical was present.
Slashdot "libertarians": Small government for me, big government for those I disagree with. -1, I disagree with you