4 replies [Last post]
phi16's picture
User offline. Last seen 16 hours 11 min ago. Offline
Joined: 12/23/2008

Every new folder, at some point, develops the urge to stretch out a new puzzle to be a long straight chain and then shake and wiggle it down into a final solution in the hopes that with only it's own internal characteristics the protein will pull itself into it's correct final state. After all, intuition tells us, putting aside divine intervention for the moment, that all of the important forces at work here are known. Of course, it never works out that way.

I've tried this more times than I'd care to admit. I've even tried stretching the amino acid chain out to a straight line; then freezing it all; then starting with #1 unfreezing while wiggling, shaking and rebuilding in hopes of recreating what it might be like to have a protein begin folding, even as it is being built, to no avail.

I can think of several reasons why not. First, I'm not so sure that all of the important forces have been accounted for. Yes, the chemistry of the amino acids are well understood; the geometry, too. But what about the surrounding environment. The water environment is all around. Since water is the third smallest molecule it must also exist between folds, as well. Moreover, the area of the cell outside the nucleus in which this is taking place is 15-20% composed of protein and protein fragments. Is any of this accounted for? I saw a very interesting YouTube animation that depicted this water environment in and around the protein molecule coming out of Croatia. I wished they had posted more.

Second, there are chaperones and chaperonin structures at work which are not well understood. I have found little information about these and would appreciate any further information you might know about out there. ONe video refers to the structure as a 'machine'. Another video called it a 'cage'. From the little information I have read, these play an important part in folding process. Fold.it might offer, as a tool much like bands, to allow folders to add a structure surrounding the protein in space which may serve to influence the folding by its 'gravity' and distance. Chaperones might also be added as a tool which simply act as a 'damper' on the amino acid chain to be added and then removed while the folding takes place.

Third, and this is more of a question than a comment, are the computations that are being made sequential? If so, can they accurately show the behavior we are looking for? In reality, each of the atoms are responding to all of the forces at work on it simultaneously. If the calculations in the Fold.it environment are done one at a time, by the time the 100th calculation for the 100th atom is made, the previous 99 atoms and their resulting positions have already shifted. Is it possible to do all of the calculations and then make the resulting shifts? This becomes a calculus problem trying to determine the resulting shift from anticipated shifts from anticipated shifts. We would need a computer for each atom, all of which were to work simultaneously. Perhaps, someone has thought to do that already.

Proteins fold very quickly. Intuitively I'd like to think that they begin folding as they are being formed. Simple ones may form secondary structures and find their final position in what would seemingly be one fluid motion. To do that, we have to find and compute all of the forces at work. We're still missing some important ones.

What do you think?

Joined: 02/24/2011
Groups: None

Starting with a fully extended chain is soooo tempting. I am struggling with it alot as yet no good results. It reminds me of the way I played golf - every shot would have been amazing if only the hole were in the right place.

Joined: 02/24/2011
Groups: None
getting past simultaneity in the simulations

Elevator pitch summary: all interaction is between the particles and the field ...

and then from the field to the particles. It's probably not a relatiivistic field. the changes that take place are likely too slow to be materially affected. but at some length (10k segments?) it starts to be an issue ...
taking the liberty of picking up a previous theme:

Think of proteins as an infection that develops and grows on a certain surface of predictability. Below that level the cost of likelihood can be quite high and short lived. The boundary is leaky. It is a fractal surface.

Look at the Mandelbrot set. Do a flyover. Catastrophies are close to equilibrium points. Holes appear from nowhere. In our solution space, phenomenologically small changes can trigger catastrophic change. I think I move it a little but it collapses. House of cards equilibrium.

Another observation. In our solution space, there are local minima/maxima everywhere. It is a extremely bumpy. What this means is that it is common for searches to get trapped. But there's another way to look at it too.

these little cups in the solution space means there are many many many configurations of the goo that are stable through shakes and wiggles. Stable. Holds its shape in the face of a reasonable range of thermal excitation (shake and wiggle). places to put things.

It is a recursive not a circular field btw.
Field(t+1, xyz) = G(Field(t, xyz)).

Joined: 08/09/2010
Groups: foldeRNA
folding as it goes

Yes indeed, proteins begin the folding process right from the beginning =)

As a side note, I usually do de novo puzzles from residue one on down the line, with decent success (I think). I use a lot of scripts that pull hydrophobes in, within a specified range, tyring to add one secondary structure at a time.

Joined: 02/24/2011
Groups: None

seems like this should be a global background that runs whenever a wiggle or shake run

I have tried it a bit but worry about the torques I am making in the strand - though perhaps those are just what is needed.

Do you use sheets as targets for the phobes?


Developed by: UW Center for Game Science, UW Institute for Protein Design, Northeastern University, Vanderbilt University Meiler Lab, UC Davis
Supported by: DARPA, NSF, NIH, HHMI, Amazon, Microsoft, Adobe, RosettaCommons