News:
Reducto ad absurdum on the LANL ASCI Q?
Posted by Jeff Perkel
[Entry posted at 25th October 2005 06:50 PM GMT]
Comment on this news story   
A research team at Los Alamos National Laboratory has topped a world record for biological simulations, and given us a taste of reducto ad absurdum.

As reported in the Nov. 1 issue of PNAS the LANL team, led by Kevin Sanbonmatsu, used 768 of the 8,192 CPUs in LANL's ASCI Q supercomputer to model the motion of 2.64 million atoms in a ribosome complex -- 250,000 in the ribosome itself, and most of the rest from water molecules. The processors chugged away for some 260 days, completing 20 million frames of motion.

Sounds impressive, right?

It is: it's six times larger than any biological simulation ever done. LANL has made a Quicktime movie of the simulation available on its Web site. The sim is a bit jerky, but it shows in atomic detail how the A-site tRNA binds to the ribosome and swings the new amino acid into position during translation, looking much like a nucleic acid crane.

According to the LANL press release, the research identified a "corridor inside the ribosome that the transfer RNA must pass through for the decoding to occur" -- a potential site for new antibiotics development. Further, the study demonstrates the active role tRNA plays in the ribosome's motion, and "sets the stage for further biochemical research".

But let's for one minute consider the numbers. The simulation's 20 million frames represent just 2 nanoseconds of real time -- or 1e-16 seconds per frame. Put another way, the computer worked at a rate of about 8 picoseconds per day. Claymation animator Nick Park's team, by contrast, produced about three usable seconds of film per day when shooting the 85-minute Wallace & Gromit: The Curse of the Were-Rabbit.

Now I'm a huge fan of tech, gadgets, and the like, and I love this study for its scale and beauty. But is this the best way to spend nine months of time on what is currently the 12th fastest supercomputer in the world? If new antibiotics really come out of this, then I suppose it was. So it's wait and see time -- just like the simulation itself.

[Via CNET News]


For FREE access to this news story and more, you must register.

Not yet registered? Get free access
 

The article you are attempting to read is only available to registered users of The Scientist. Registration is FREE and only takes a few seconds.

 
 

Email

Password

> Forgot Password?
> FAQ
> Subscribe

 
Not yet registered? Get free access
 

Create your MyScientist account and access all of The Scientist's free content, tools and life science email newsletters, including:

 

> The current month’s print issue

> Daily & Bi-weekly e-mail newsletters

> Newsblogs with breaking headlines

> The Scientist Community

> Exclusive web extras

> The Scientist Careers

 

Premium content from The Scientist Archive, a comprehensive resource of over 22 years of past life science coverage, is available only by subscription. Subscribe today and get unlimited access

 

 
LATEST NEWS