Biological Computer Created at Stanford

sciencehabit writes "For the first time, synthetic biologists have created a genetic device that mimics one of the widgets on which all of modern electronics is based, the three-terminal transistor. Like standard electronic transistors, the new biological transistor is expected to work in many different biological circuit designs. This should make it easier for scientists to program cells to do everything from monitor pollutants and the progression of disease to turning on the output of medicines and biofuels."

Comments

Simulated vs. Real results

[girlinatrainingbra]

There's a good picture of the "simulated results" vs. the results they really got in that Science magazine preview for an AND gate, and a relevant paragraph of the summary : http://news.sciencemag.org/sciencenow/assets/2013/03/28/sn-circuit.jpg
The Stanford team then showed that they could line up multiple transcriptors to carry out logical functions, creating standard logical circuits called AND gates, OR gates, XOR gates, and so on, which combine signals according to certain rules. (A computer's processor is a vast assemblage of such gates.) They also showed that their novel biological circuit designs were adept at producing signals with large amplification and that they could be used to up the expression of a variety of genes, such as the production of fluorescent signals that made it simple to detect cells that were carrying out their programming.

I wonder exactly how they "assemble" the circuit and keep the components from diffusing or floating away, thus diassembling the circuit. What keeps the "circuit" of DNA strands in place?

Re:Simulated vs. Real results

[Biotech_is_Godzilla]

Read TFA for the components of this circuit. The DNA part of the circuit is most likely integrated into the cellular genome, so is effectively stationary in the nucleus. The RNA polymerase component is probably the naturally occurring version of the protein that already exists in the cell. RNA polymerase randomly diffuses around in the nucleus, but there's not just one molecule of RNA polymerase around, there's loads of them, and they can all do the same job. With help from other proteins, they bind to sequences in the genomic DNA that mean "make this gene under this condition", and transcribe the DNA into RNA, which then gets made into protein.

The conditions under which RNAP binds and transcribes can be dependent on the cell receiving certain signals from other cells, or a gene may be transcribed 'constitutively', meaning it is always transcribed unless the cell changes state/gets a signal from other cells. That's probably what they're using here - RNAP will transcribe what it can transcribe under all conditions, its just that you change what it's allowed to transcribe (see below). The odds of RNAP hitting the DNA part of the circuit are going to be high, and once they do hit the 'promoter' sequence in the DNA part of the circuit, they will lock onto it and start transcribing. They can't transcribe the interesting bit (the 'signal') unless it has been switched on by the third component of the 'circuit', integrase, which removes (and puts back) the "STOP" control for transcription (by cutting at defined sequences, specific to the integrase, either side of the STOP control).

So the integrase is acting as the gate, and the 'signal' represents electrons flowing to the drain... or whatever. IANACS.

The method of control for these circuits is probably on the level of the whole cell - the researchers will be adding signalling chemicals to the cells that switch on production of / alter the cellular location of the integrase so that it can either do its job or not. The rest of the circuit doesn't have to be directly controlled as it will constitutively do its job.

The potentially interesting bit comes from making the 'signal' that's produced be the chemical signalling molecule that controls expression/localisation of integrase in other cells. The problem they'd then immediately run into is that you can't stop the signalling molecule from also acting on the cell that's generating it, locking it into an "on" state, unless the other cells use a different signalling molecule to control their integrase/ use integrases that act at a different specific defined sequence. They could do that, but it would then be difficult to control the placement of cells containing the different flavours of transistor.

This is all speculation as I haven't got access to the full article, but from the abstract I'm fairly confident that's what these guys have done. It's not about to lead to anything remotely resembling a proper computer any time soon. As I biologist I automatically think of anything associated with 'synthetic biology' as being sensationalist rubbish, and I don't think this is an exception, sadly.

Good point, but a dozen can run vmware

[raymorris]

Perhaps the headline should have said "logic gates" instead of "computer". It didn't say "Core i7" either, though. Babbage's machine was a computer. Programing graphics processors with punch cards dates to the early 1800s, so "computer" doesn't imply a modern desktop.

I suspect you'd agree that any processor capable of running Windows is a computer. Therefore, any machine that can run a hypervisor, which in turn runs Windows, is a computer. You probably know where I'm headed - Turing machines. Any Turing machine can emulate a Core processor, and is therefore a computer. Wolfram's Turing machine requires only a few gates, so these researchers can probably build a biological Wolfram Turing computer today.