Folding Forum

http://www.iflscience.com/technology/wo ... mes-second

I wonder if we could soon have a recording of an actual protein folding?

TonyStewart14 wrote:http://www.iflscience.com/technology/wo ... mes-second

I wonder if we could soon have a recording of an actual protein folding?

I posted that link in another forum I use, and someone commented:

And you still wouldn't be able to capture the moment that a political party u-turns on policy...

But being serious- as I understand it, it uses light in the (more or less) visible range, which has far too long a wavelength to be of any use for watching atoms move.

You'd need to go to X-rays or beyond, and promptly run into the brick wall of the observer effect, and anyway it's too slow by a couple of orders of magnitude (using the FAH step size of 2 femto-seconds as a guide)

According to wikipedia....
1.3 fs – cycle time for 390 nanometer light, at the transition between violet visible light and ultraviolet[5]
2.57 fs – cycle time for 770 nanometer light, at the transition between red visible light and near-infrared[5]

and this camera does a frame every 2.2^-13 seconds (a fs is 10^-15 seconds)
so quite close but no cigar...

http://en.wikipedia.org/wiki/Femtosecond

Bob

popandbob wrote:According to wikipedia....
1.3 fs – cycle time for 390 nanometer light, at the transition between violet visible light and ultraviolet[5]
2.57 fs – cycle time for 770 nanometer light, at the transition between red visible light and near-infrared[5]

The cycle time (reciprocal of frequency) isn't the problem- also from Wiki, bond lengths and atomic diameters are in the range 100-200 picometers so the spatial resolution is insufficient by a factor of at least 3,000, even for light at the blue end of the spectrum.

Someone from PG might be better qualified to comment than I am (it's not my area of expertise) but I doubt that, using visible light, it would be possible to resolve even a single giant protein molecule, let alone the small ones that FAH can handle. Those wishing to study molecular structure use X-ray crystallography for good reasons, and that's a technique about as unsuited to producing real-time movies as you can get

popandbob wrote: and this camera does a frame every 2.2^-13 seconds (a fs is 10^-15 seconds)
so quite close but no cigar...

http://en.wikipedia.org/wiki/Femtosecond

Yup, a factor of 110 more than FAH's step time. Opinions vary of course, but I don't call that "quite close"

I always hesitate to say that anything technological is truly impossible, but a movie of a protein folding seems pretty unlikely. That having been said, I've often wondered if PG have ever used their data to produce a "movie" of a trajectory from start to finish, it could be a useful "advert" for the FAH project.

"Start to finish" data has not yet been completed, even for relatively small proteins.

THe most popular video shows one full millisecond of folding of NTL9.
https://www.youtube.com/watch?v=gFcp2Xpd29I
The video was(produced a few years ago so you may have seen it. Work is still continuing at a relatively low priority.

Bond vibration takes place on the femtosecond time-scale. Many processes take place at pico- nano- micro- or milli-second time-scales, and some even take as much as a full second. That's a lot of steps to simulate.

bruce wrote:"Start to finish" data has not yet been completed, even for relatively small proteins.

That's certainly a valid enough objection

bruce wrote:THe most popular video shows one full millisecond of folding of NTL9.
https://www.youtube.com/watch?v=gFcp2Xpd29I
The video was(produced a few years ago so you may have seen it. Work is still continuing at a relatively low priority.

I hadn't seen it before, thanks for the link. Exactly the sort of thing I was thinking of, and bookmarked for future reference.

bruce wrote: Bond vibration takes place on the femtosecond time-scale. Many processes take place at pico- nano- micro- or milli-second time-scales, and some even take as much as a full second. That's a lot of steps to simulate.

And a lot of video frames to stick together… or, more realistically, deciding which 99.9999% to leave out! I'd guess that the trajectory can't just be sampled at fixed intervals as there may be long periods where little of interest (from a "publicity" point of view) happens interspersed with bursts of activity as parts of the molecule move into critical positions?

You're right. One of the primary problems that FAH had to solve before useful results were obtained was what you call "stick together" Folding has short periods of "events" when a lot happens but they get lost in long periods of "dwell" where bonds just oscillate back and forth making uninteresting motions.

We're desperately off-topic here.

bruce wrote: We're desperately off-topic here.

Yes, sorry about that. But you've told me what I wanted to know so I'm happy to leave it there.

I have no doubt, that if the capability, precision and accuracy existed then someone would write some software to compare images and thereby have the ability to have a computer condense the unneeded and/or surplus information in the video to more usable levels.

My opinion: Really, are we that far topic? Perhaps it has strayed a bit but there is still a common string attached and it has become more pertinent to folding as it has morphed.

How about this TED talk? http://www.ted.com/talks/ramesh_raskar_ ... per_second
Mirror video: https://www.youtube.com/watch?v=Y_9vd4HWlVA