Previous Thread
Next Thread
Print Thread
rece vs unfold
#18074 05/26/08 04:31 PM
Joined: Sep 2003
Posts: 250
jb007 Offline OP
Forum Member
OP Offline
Forum Member
Joined: Sep 2003
Posts: 250
I have a simple(?) question.

Post processing MD trajectories created with PBC offers either the recenter command (in which the selected atoms are fitted into the primary cell as far as I understood it). However, an alternate is the unfold command, allowing the molecule moving w/o forcing it back to the primary unit cell.

My question is: Shouldn't the unfold option - here allowed for all atoms - reflect the real event by means of solute, solvent molecules (ions, water etc) remaining bound at the first hydration layer (2.4 A)? A comparison between the rece and unfold options results in completely different behavior; the latter exhibit a decay that levels of. I attach a graph showing what I mean for the unfold case.

Thanks in advance for an eventual answer.

best,

Johan

Attached Files
18140-Graph1.jpg (0 Bytes, 176 downloads)
Re: rece vs unfold
jb007 #18075 05/26/08 05:39 PM
Joined: Sep 2003
Posts: 8,447
rmv Online Content
Forum Member
Online Content
Forum Member
Joined: Sep 2003
Posts: 8,447
Experiments have shown that the residence time for an individual water in the solvation shell is not all that long, on the order of a few ps on average. There is a fairly rapid exchange between bulk and solvation shell waters.

UNFOLD removes the effects of image centering, so that the atomic coordinates are continuous (no unit cell jumps). In general, a water box becomes fairly diffuse from this operation. The primary use has been for estimates of diffusion. Many of the waters interacting with the protein will be outside the box after the unfolding.

Re: rece vs unfold
rmv #18076 05/26/08 06:08 PM
Joined: Sep 2003
Posts: 250
jb007 Offline OP
Forum Member
OP Offline
Forum Member
Joined: Sep 2003
Posts: 250
thanks,
this would imply that solvent re-centering may not actually capture the real hydration pattern in - say - the first solvation layer around the solute when evolving in time? Or am I missing something fundamental here? Given that the water at some points would stick with the protein during the diffusion, which I assume is random leveling of to some constant hydration? I let all atoms unfold in this test and did a similar thing computing the rdf of sodium around DNA phosphate oxygens. The Na+-O rdf curvature for the unfold processed trajectory is very similar to the re-centered, let alone the first peak in the recenter case being slightly higher.

Re: rece vs unfold
jb007 #18077 05/26/08 06:21 PM
Joined: Sep 2003
Posts: 8,447
rmv Online Content
Forum Member
Online Content
Forum Member
Joined: Sep 2003
Posts: 8,447
RECEnter should preserve the solvent and ion packing from the simulation, at the expense of translation (and solute rotation if ORIENT was used).

There's no such guarantee for UNFOLD (quite the contrary).

Ion pairs do tend to last a good deal longer than water molecule interactions. I would expect different behavior.

Re: rece vs unfold
rmv #18078 05/26/08 07:20 PM
Joined: Sep 2003
Posts: 250
jb007 Offline OP
Forum Member
OP Offline
Forum Member
Joined: Sep 2003
Posts: 250
Rick,
you've made things crystal clear

I appreciate your answers and conclude that the leveling of in waters actually might be coordinated to Na+ ions. Of course I might be wrong, but that won't be the first time. Or last time either for that matter, thanks


Moderated by  John Legato, lennart 

Link Copied to Clipboard
Powered by UBB.threads™ PHP Forum Software 7.7.4
(Release build 20200307)
Responsive Width:

PHP: 5.6.33-0+deb8u1 Page Time: 0.013s Queries: 25 (0.006s) Memory: 0.9227 MB (Peak: 1.0162 MB) Data Comp: Off Server Time: 2020-05-28 18:02:29 UTC
Valid HTML 5 and Valid CSS