Building a DNA Computer?
Thread - DNA Computing
On 27/2/2003, Toby Fiander
wrote:
The ABC Science site is carrying an article on DNA computing. The nub of
the story is that a an Israeli lab has invented "the first programmable autonomous
computing machine in which the input, output, software and hardware were
all made of DNA molecules. It can perform a billion mathematical operations
a second with 99.8% accuracy."
... and.....
"Most importantly, the new design incorporates a previously unknown biochemical
process that generates enough heat energy to power the device, meaning in
principle that a DNA computer can work without an external energy source."
I suppose I will be sorry for asking, since I have little understanding beyond
first principles of how a conventional computer works, but could anyone indicate
what advantage a DNA computer might be and, I suppose, a bonus question,
how does it actually operate?
Will the keyboard be made of gelatine? Is this another case of biology
meets engineering, Moore or less?
Chris
Lawson replied:
>"Most
importantly, the new design incorporates a previously
>unknown biochemical process that generates enough heat energy to
>power the device, meaning in principle that a DNA computer can
>work without an external energy source."
I'm sorry, but
now we not only have a working DNA computer, but a perpetual motion machine
as well? This is either a con job or a very bad piece of reporting.
Zero Sum added:
Humm... Where to start... Well firstly, what they have done is
produce a new "engine of computation". To understand that we will have
to explain a few first principles. I'm sure we all know them, but I
need to string a few beads.
It is possible to build a device that can read two switches and set a third
switch depending on the state of the first two switches. There are
a number of types of these devices that can "AND", "OR" or "XOR" (etc.) the
first two switches and set the third switch accordingly. These devices
can be implemented mechanically, electrically, with billiard balls on a billiard
table, DNA or almost anything you care to think of. They are simple
devices, it doesn't take much to make them. I am sure you could do
it with pipes and water pressure. Now, I think it was in the late 80's
that they worked out how to do it with DNA. But it isn't really saying
much.
The simple devices we are talking about are used to make larger more complex
devices. Some are devices entire, such as a moisture measure but more
importantly for computing, the simple devices we mentioned can be wired together
to make something like a "binary adder" which adds two numbers stored in
an array of switches and sets another array of switches with the result.
These and other devices might be strung together to make some more complex
device (that we might call an "instrument") such as an early modem.
But we are still talking "non programmable", "not a computer" at this level.
When you build these devices together and allow them to follow a set of operations
stored in more switches, you start to look like the early days of computers.
But what you build your computer of does not matter to the computer.
The computer is an abstract that is based only on those original simple devices
strung together. It will run in any form.
So, just having built the basic primitives, well, that's the easy
part.
Turning this into a usable computer is a long way away.
Paul
Williams added:
<snip>
> I suppose I will be sorry for asking, since I have little
> understanding beyond first principles of how a conventional
> computer works, but could anyone indicate what advantage a DNA
> computer might be and, I suppose, a bonus question, how does it
> actually operate?
>
I believe that
the huge advantage is that instead of performing one task at a time in series,
they can perform multiple tasks at the same time (parallel processing).
"They
are massively parallel, compute with extremely high energy-efficiency and
store enormous quantities of information."*
- Leonard
Adleman
*I found
this article from last year helpful:
http://www.jpl.nasa.gov/releases/2002/release_2002_63.html
