Suddenly Greg Bear's Blood Music seems less fanciful and more than a little worryingly real...
Boffins make bio-chip breakthrough
MIT boffins have made a breakthrough in biological computing that paves the way for cancer-detecting yogurts and other gloopy marvels. The advance, which saw the researchers combine logic and memory within a single living cell, was published in the "Synthetic circuits integrating logic and memory in living cells" paper in …
-
Wednesday 13th February 2013 19:14 GMT frank ly
re. "Messy collection of data aside, ..."
You're not thinking this through. Combine this breakthrough with DNA from electric eels and fireflies, then the bacterial sensors will be able to send data via bluetooth or by encoded flashes as they lay steaming in the pan.
The possibilities are endless and potentially very entertaining (your sense of humour may vary).
-
Wednesday 13th February 2013 19:42 GMT John Smith 19
Impressive.
So we have logic + a data tape.
Need a system that reads the DNA and using it to configure the logic gates to accept inputs and process them.
A sort of "program" if you will.
True it's unlikely to break speed records.
But check the density. Less a processor array and rather more "smart matter," as people talked about the arrays to do CFD simulation.
Thumbs up. This may be one of the ways of the future.
-
Wednesday 13th February 2013 21:49 GMT MondoMan
Re: Impressive.
The density seems to be 4 bits in a cubic micron (assuming a 2D layer of cells, that would be a feature size of 250 nm -- not too impressive).
There's already a perfectly good processing system in the cell; trying to crudely implement boolean logic in DNA seems about as elegant and useful as simulating CPU function on sheets of paper with a pencil.
-
This post has been deleted by its author
-
Saturday 16th February 2013 16:43 GMT John Smith 19
Re: Impressive.
"he density seems to be 4 bits in a cubic micron (assuming a 2D layer of cells, that would be a feature size of 250 nm -- not too impressive)"
I'll make 2 points.
How many layers of those memory elements can you lay down on a semiconductor chip, not material layers. Actually unified working components?
Because that's pretty much 1 layer (3d chips with multiple layers of logic and/or memory have not worked out well), versus as many as you like with a protein/DNA system. Those DNA layers are likely to be 10-100x thinner than any viable semiconductor layer (which on conventional wafers sits on about a 300-500 micrometre thick silicon layer).
Secondly this is v0.1 technology. It's not got data input and to make proper use of existing data mechanisms (the ribosome) you'd want to store in codons and a double helix with 1 twist giving six bits of storage.
DNA give you volumetric density in way conventional semiconductor processing simply does not.
Changing the mindset to play to its strengths and adapt to its weaknesses will be challenging.
-
-
-
-