From: Francesco Pietra (chiendarret_at_gmail.com)
Date: Tue May 24 2016 - 04:30:13 CDT
>
> (1) But I suspect Drude (or another polarizable force field) is the only
>> possible solution for what you want.
>>
>> (2) Or you have to ask less precise questions about the system.
>>
>
Option (1) could help about H-bonds. However, neither Drude, nor other
polarizable ff, are compatible with metalloproptreins of d-block transition
metals, which is my problem besides H-bonds.
Option (2) is what I can stick to. Which means that with metalloproteins
of d-block transition metals, the year 2016 is still Middle Age as far as
classical MD is concerned.
thanks a lot or all your precious comments and suggestions
francesco pietra
On Mon, May 23, 2016 at 3:59 PM, JC Gumbart <gumbart_at_physics.gatech.edu>
wrote:
> Whatever angle psfgen sets the H at is determined by (1) internal
> coordinates in the topology file (the IC lines) and (2) some simple
> heuristics. That angle should not be considered final, but just a first
> guess; run a minimization to see what the resulting *preferred* angle is.
>
> If I understand you correctly, you’re talking about fixing the angles to
> what they should be in simulation, but all that does is push the inaccuracy
> in the force field somewhere else. I don’t think you can claim it gives
> you any particular insight.
>
> I have no idea about future psfgen support for Drude - we use charmm tools
> (including charmmgui.org) to prepare files for those simulations. But I
> suspect Drude (or another polarizable force field) is the only possible
> solution for what you want. Or you have to ask less precise questions
> about the system.
>
> Best,
> JC
>
> On May 22, 2016, at 2:46 AM, Francesco Pietra <chiendarret_at_gmail.com>
> wrote:
>
> Sorry, replace "in my case 11 instead of 180" with "in my case 111 instead
> of 180"
> ---------- Forwarded message ----------
> From: Francesco Pietra <chiendarret_at_gmail.com>
> Date: Sun, May 22, 2016 at 8:43 AM
> Subject: Re: namd-l: H-bonds in charmm36 with namd2.11
> To: JC Gumbart <gumbart_at_physics.gatech.edu>
> Cc: NAMD <namd-l_at_ks.uiuc.edu>
>
>
> What about setting distance and angle colvars (with namd2.11/charmm36)
> about key H-bonds, as determined by H-ray diffraction, and built on psfgen
> (psfgen fails to set the angle correctly, in my case 11 instead of 180
> degrees)? In order to study what happens elsewhere during MD.
>
> The problem is most acute with metalloproteins, with the active center
> built around d-block transition metals. Which is my case. Currently
> hopeless for me using polarized ff, although my model for the metal ion is
> polarized, though not in the frame of available polarized ff.
>
> Incidentally, is any future prospect for psfgen with Drude or other
> polarized potential?
>
> Thanks
>
> francesco pietra
>
> On Sun, May 22, 2016 at 1:44 AM, JC Gumbart <gumbart_at_physics.gatech.edu>
> wrote:
>
>> There’s a few factors at play here, the main one being lack of
>> polarization. Beta sheets in particular can not reproduce the correct
>> geometry of hydrogen bonds in a fixed-charge force field (at least that’s
>> what I recall).
>>
>> Here’s a recent paper that seems especially relevant:
>> http://pubs.acs.org/doi/abs/10.1021/jp503972j
>>
>> Best,
>> JC
>>
>>
>> On May 20, 2016, at 5:18 PM, Francesco Pietra <chiendarret_at_gmail.com>
>> wrote:
>>
>> Hello:
>>
>>
>> H-bonds in the charmm-ff are substantially viewed as in the excerpts
>> below:
>>
>>>
>>> *From:* Tim Isgro (*timisgro_at_ks.uiuc.edu*
>>> <timisgro_at_ks.uiuc.edu?Subject=Re:%20%20hydrogen%20bonds%20in%20NAMD>)
>>> *Date:* Fri Mar 19 2004 - 11:25:47 CST
>>>
>>> - *Next message:* Streiff, John H.: "initial temp of simulation"
>>> <http://www.ks.uiuc.edu/Research/namd/mailing_list/namd-l.2003-2004/0720.html>
>>> - *Previous message:* Stern, Julie: "running multiple minimizations"
>>> <http://www.ks.uiuc.edu/Research/namd/mailing_list/namd-l.2003-2004/0718.html>
>>> - *In reply to:* Ioana Cozmuta: "hydrogen bonds in NAMD"
>>> <http://www.ks.uiuc.edu/Research/namd/mailing_list/namd-l.2003-2004/0710.html>
>>> - *Messages sorted by:* [ date ]
>>> <http://www.ks.uiuc.edu/Research/namd/mailing_list/namd-l.2003-2004/date.html#719> [
>>> thread ]
>>> <http://www.ks.uiuc.edu/Research/namd/mailing_list/namd-l.2003-2004/index.html#719> [
>>> subject ]
>>> <http://www.ks.uiuc.edu/Research/namd/mailing_list/namd-l.2003-2004/subject.html#719> [
>>> author ]
>>> <http://www.ks.uiuc.edu/Research/namd/mailing_list/namd-l.2003-2004/author.html#719> [
>>> attachment ]
>>> <http://www.ks.uiuc.edu/Research/namd/mailing_list/namd-l.2003-2004/attachment.html>
>>>
>>> Hi Ioana,
>>> The hydrogen bond potential in the Charmm force field is, from what I
>>> gather, viewed skeptically by many people, since there is no very good
>>> way
>>> to benchmark the parameters and potential.
>>>
>>> *From:* Axel Kohlmeyer (*akohlmey_at_gmail.com*
>>> <akohlmey_at_gmail.com?Subject=Re:%20%20question%20on%20hydrogen%20bond>
>>> )
>>> *Date:* Mon Sep 06 2010 - 04:08:32 CDT
>>>
>>> - *Next message:* Ramya Narasimhan: "user-defined energy function"
>>> <http://www.ks.uiuc.edu/Research/namd/mailing_list/namd-l.2009-2010/3673.html>
>>> - *Previous message:* dhacademic: "Re: FATAL ERROR: Due to a design
>>> error, GlobalMasterServer does not support individual atom requests from
>>> multiple global force clients on parallel runs."
>>> <http://www.ks.uiuc.edu/Research/namd/mailing_list/namd-l.2009-2010/3671.html>
>>> - *In reply to:* Kwee Hong: "question on hydrogen bond"
>>> <http://www.ks.uiuc.edu/Research/namd/mailing_list/namd-l.2009-2010/3670.html>
>>> - *Next in thread:* ehenriques_at_qui.uc.pt: "Re: question on
>>> hydrogen bond"
>>> <http://www.ks.uiuc.edu/Research/namd/mailing_list/namd-l.2009-2010/3674.html>
>>> - *Messages sorted by:* [ date ]
>>> <http://www.ks.uiuc.edu/Research/namd/mailing_list/namd-l.2009-2010/date.html#3672> [
>>> thread ]
>>> <http://www.ks.uiuc.edu/Research/namd/mailing_list/namd-l.2009-2010/index.html#3672> [
>>> subject ]
>>> <http://www.ks.uiuc.edu/Research/namd/mailing_list/namd-l.2009-2010/subject.html#3672> [
>>> author ]
>>> <http://www.ks.uiuc.edu/Research/namd/mailing_list/namd-l.2009-2010/author.html#3672> [
>>> attachment ]
>>> <http://www.ks.uiuc.edu/Research/namd/mailing_list/namd-l.2009-2010/attachment.html>
>>>
>>> On Sun, Sep 5, 2010 at 11:40 PM, Kwee Hong <joyssstan0202_at_gmail.com
>>> <joyssstan0202_at_gmail.com?Subject=Re:%20%20question%20on%20hydrogen%20bond>>
>>> wrote:
>>> *> Hi. *
>>>
>>> joyce,
>>>
>>> *> In the namd-tutorial-unix.pdf, it mentioned that explicit hydrogen
>>> bond *
>>> *> terms are no longer present in the CHARMM force field and are
>>> therefore not *
>>> *> calculated by NAMD. Does this mean that throughout the simulation,
>>> hydrogen *
>>> *> bonding are not calculated? So does that mean that I can't inspect
>>> the *
>>> hydrogen bonds are a phenomenological description and thus
>>> technically a not an easily identifiable term for a force field.
>>> they are more a "side effect" of the charge distribution and
>>> geometry of groups involving hydrogens bound to certain atoms
>>> (h-bond donors) and their counter parts (h-bond acceptors
>>>
>>
>>
>>
>> Be that as it may be, is charmm36, as run with namd2.11 for a protein in
>> TIP3 water solution, capable of maintaining all distances between aa
>> O-atoms with interposed H atom, as determined by X-ray diffraction of the
>> protein crystal? When not (as I am observing), should one suspect that the
>> fictitious charge distribution was incorrectly calculated? Or simply assume
>> that crystal and water solution are different affairs? Which would imply
>> that crystallographers should refrain from making too many conclusions from
>> their structural data.
>>
>> Grateful for a recent point of view.
>>
>> francesco pietra
>>
>>
>>
>
>
>
This archive was generated by hypermail 2.1.6 : Tue Dec 27 2016 - 23:22:11 CST