Re: Excluding interactions?

From: Jérôme Hénin (
Date: Thu Jan 21 2010 - 08:17:26 CST

Oops. Indeed it sounds like it should always happen, but I remember
doing something similar, where at least a few atoms were far apart yet
had nonbonded exclusions, and NAMD managed it just fine. That was
probably around NAMD 2.5, though: it is possible that this behavior
has changed since then.
Sorry I can't be any more help...

2010/1/20 Charles Zhao <>:
> Now that I have done this, I am unable to start a simulation without it
> crashing due to a "bad global exclusion count".  Reading the line here:
> , it says "In rare
> special circumstances atoms that are involved in bonded terms (bonds,
> angles, dihedrals, or impropers) or nonbonded exclusions (especially
> implicit exclusions due to bonds) will be placed on non-neighboring patches
> because they are more than the cutoff distance apart. This can result in the
> simulation dying with a message of ``bad global exclusion count'', my
> interpretation is that this will always happen if I attempt something like
> this.  Is it at all possible to avoid?
> ________________________________
> From: Jérôme Hénin <>
> To: Charles Zhao <>
> Cc:
> Sent: Sat, January 16, 2010 10:37:26 AM
> Subject: Re: namd-l: Excluding interactions?
> That is correct, if you want to involve more than two groups,
> Alchemify won't do it for you. If you can do a bit of programming
> though, I recommend to try and adapt the Alchemify code to do exactly
> what you want. It is probably the easiest way.
> Jerome
> 2010/1/16 Charles Zhao <>:
>> Thank you for your response.  I think I understand.  With this I can, for
>> instance, remove nonbonded interactions between A and B in a psf file.
>> However, if I then remove the interactions between C and A using the same
>> procedure, will Alchemify keep both nonbonded exclusion lists, or erase
>> the
>> first one between A and B?  I would of course like to have both at once.
>> The web page is not clear on this, and it seems Alchemify can only remove
>> the interactions between two groups at once (since that's what it was
>> designed for).
>> ________________________________
>> From: Jérôme Hénin <>
>> To: Charles Zhao <>
>> Cc:
>> Sent: Fri, January 15, 2010 4:53:51 AM
>> Subject: Re: namd-l: Excluding interactions?
>> Hi,
>> They were probably referring to nonbonded exclusions (originally a
>> CHARMM feature). They can be specified by a NNB section at the end of
>> the PSF file. One convenient way to build such a section to exclude
>> interactions between two groups is to use alchemify
>> (, pretending that each of the
>> groups is an alchemical end-point state. The web page has more info.
>> Jerome
>> 2010/1/15 Charles Zhao <>:
>>> I have heard from some people I have spoken to that it is possible to
>>> convince namd to ignore interactions between certain groups of atoms in a
>>> system (i.e. between peptide B and C, but not between B and A or C and
>>> A).
>>> This would in fact be quite useful to me.  However, it seems like a vague
>>> idea, and they themselves do not remember clearly how it is done.  I have
>>> not been able to find any documentation regarding anything like this.  Is
>>> this indeed possible, or am I mistaken?

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:55:21 CST