AW: AW: Periodic Box Dimension (WrapAll is on)

From: Norman Geist (
Date: Tue Jan 20 2015 - 02:07:47 CST

Gesendet: Dienstag, 20. Januar 2015 08:37
An:; Norman Geist
Betreff: Re: AW: namd-l: Periodic Box Dimension (WrapAll is on)




Thank you for the comments,


Based on my humble knowledge, when an atom crosses the boundry should enter from the other side. So,during an MD with pbc , all atomic coordinates in the dcd file should lay within the box dimension.


No, not by default.


However, VMD showes some atoms sticked out of the box. I know that VMD does not do magic , simply reads coordinates from NAMD dcd and display them.


I can not figure out why NAMD produce atomic coordinates which lay out of the box.


I know , i am doing some thing stupid, would you mind to clarify on " wrapping is only a “output” option"?


NAMD doesn’t wrap the atoms on DCD output if you do not tell it to by f.i. “wrapAll” or “wrapNearest”. But I and many other do not wrap in NAMD as it wraps every single atom which looks terrible and produces long bonds crossing the cell f.i. in water. You can produce the same behavior in VMD by entering into console:


pbc wrap -all


You will not like the result for “watching”, but this kind of wrapping might be required when doing analysis where the atoms really have to at the places, where they actually are, like density. But for visualization purpose one usually uses:


pbc wrap -all -compound res -center com -centersel protein


Please see the separate PBC tools manual for details. Try both commands and see what happens (You will need to reload your DCD after the 1st command as it is hard to get rid of the long bonds again). Be careful when dealing with pbc post processing.


By the way, i am an chem eng and new to MD technique.




From: Norman Geist <>
Sent: Tuesday, 20 January 2015, 10:09:59
Subject: AW: namd-l: Periodic Box Dimension (WrapAll is on)



Von: [] Im Auftrag von JAVAD NOROOZI
Gesendet: Dienstag, 20. Januar 2015 06:55
An: NAMD list
Betreff: namd-l: Periodic Box Dimension (WrapAll is on)






As far as i know NAMD does not wrap atomic coordinates to the other side of the box until the center of mass of the molecule composed of those atoms pass the boundary.


well as far as I know NAMD wrapps _EVERY_ atom that crosses the boundary, when wrapping is enabled. As this is often not comfortable, people usually turn of wrapping in NAMD and wrap later in VMD with much more options.


The question is what would happen if one specify the periodic box dimension much smaller than the max-min of the initial coordinate file(#.coord)?


In this case center of mass of some molecules stick out the box, does NAMD wrap them to the other side in the very first step?


What if some atoms in a molecule stick out (not the molecule center of mass ) of box?


What happens is that your simulation will very likely crash because of superimposed atoms and very high VDW energy, as for those atoms sticking out of the box now, there’s no space on the other side of the box. Please also notice that wrapping is only a “output” option. The simulation itself will always handle the atoms periodic, if a box is set at all. Seems that you have some critical misunderstanding in some of the fundamentals of molecular dynamics simulation.


What are you trying to do?








This archive was generated by hypermail 2.1.6 : Tue Dec 27 2016 - 23:20:49 CST