Re: ion moving laterally during Umbrella Sampling?

From: Giacomo Fiorin (giacomo.fiorin_at_gmail.com)
Date: Wed Oct 04 2017 - 12:05:39 CDT

On Wed, Oct 4, 2017 at 12:48 PM, Olya Kravchenko <ovkrav_at_gmail.com> wrote:

> Thank for replying!
>
> I have put both colvars into one *.in file and I can see that the output
> shows both colvars in one output file. I am not sure how WHAM interprets
> the output file though, since it contains additional column of data.
>

That depends on the WHAM implementation but if you are not sure you can
write only the columns of interest using a script, or even just run this
one:

https://raw.githubusercontent.com/Colvars/colvars/master/
colvartools/plot_colvars_traj.py

> >>Keep in mind that if you put restraints that affect the ion conduction
> pathway, you have to question yourself whether you are changing the
> resulting PMF, which has to assume a minimum free energy pathway.
>
> During regular equilibration (not US runs) I do observe interaction
> between ions and this particular site, however ions don't stay there long,
> they proceed almost immediately further down the channel towards stronger
> sites inside. I think that because the ion is confined to move in the small
> window it shifts towards this residue in this particular window (and a
> couple of adjacent windows) and does not leave for the rest of the run
> because interaction with the residue is too strong. Does it make sense?
>

If you think this behavior is specific to your system, I cannot say much.

In a general case, be aware that if your minimum free-energy path is a
curve in a higher space than the one you have chosen (i.e. distanceZ only),
biasing forces can try too hard to push over barriers that are "hidden"
because of the additional variables that you are not biasing. Restraining
such additional variables may fix histeresis problems, but at the cost of
artificially changing the path.

>
> I wonder if increasing the force constant in the input file would affect
> this? I use 2.5 kcal/mol*A with width 1.0, and it works well inside the
> channel. Is it reasonable to use a bigger force constant (say, 5 or
> 7kcal/mol*A) at the entrance or is it better to have the same force
> constant for each window during PMF simulations?
>

This is most certainly a question for someone who is an expert in your
system.

>
> Olga
>
>
>
>
>
>
> On Tue, Oct 3, 2017 at 4:16 PM, Giacomo Fiorin <giacomo.fiorin_at_gmail.com>
> wrote:
>
>> You can split configuration for different variables in multiple files,
>> and load them using "cv configfile" or even put them as strings (between
>> double quotes) in the NAMD main script and load them through "cv config".
>>
>> Keep in mind that if you put restraints that affect the ion conduction
>> pathway, you have to question yourself whether you are changing the
>> resulting PMF, which has to assume a minimum free energy pathway.
>>
>> Giacomo
>>
>> On Tue, Oct 3, 2017 at 1:45 PM, Olya Kravchenko <ovkrav_at_gmail.com> wrote:
>>
>>> Hi all,
>>>
>>> in my current umbrella sampling simulations for ion permeation through
>>> the channel I see ion shifting off the center line (where it supposed to be
>>> located) near the entrance or exit of the pore (no shift in the windows
>>> inside the pore). It seems to move to the nearest strong residue and stay
>>> there throughout simulation without returning back to the specified centre.
>>> The ion can move 2-3 A to the side over the first 0.5-1 ns and linger there
>>> for the rest of the simulation. I use distance colvar for this simulation.
>>>
>>> I would expect that for umbrella sampling the ion should not shift that
>>> far from a specified location (and I haven't seen this in my previous US
>>> simulations on a similar system). What would be the reason for this to
>>> happen in US simulations?
>>>
>>> I want to try using distanceXY colvar to restrain this movement. Is is
>>> applicable in this case, in addition to distance colvar (i.e., not
>>> distanceZ, but distance)? Do I have to have a separate input file for
>>> distanceXY colvar or can I add it into the same *.in file as I have for
>>> distance colvar?
>>>
>>> Thanks in advance, I appreciate your advice.
>>>
>>> Olga
>>>
>>>
>>>
>>>
>>
>>
>> --
>> Giacomo Fiorin
>> Associate Professor of Research, Temple University, Philadelphia, PA
>> Contractor, National Institutes of Health, Bethesda, MD
>> http://goo.gl/Q3TBQU
>> https://github.com/giacomofiorin
>>
>
>

-- 
Giacomo Fiorin
Associate Professor of Research, Temple University, Philadelphia, PA
Contractor, National Institutes of Health, Bethesda, MD
http://goo.gl/Q3TBQU
https://github.com/giacomofiorin

This archive was generated by hypermail 2.1.6 : Sun Dec 31 2017 - 23:21:41 CST