[gmx-users] 5. Re: 3. Re: Strange pullx coordinates (PMF calculations)

[Justin Lemkul]

On 11/5/18 5:07 AM, CROUZY Serge 119222 wrote:
> Hello Justin-
> In MY pullx first column is Time and second column is absolute coordinate of the COM of the pulled group
> Maybe we are missing an option which would print X and dX in the pullx files - one of the pull-print stuffs ???!!.. In that case too bad we would have tons of "bad" pull files Printing the reaction coordinate (dX) should be the default .. Don't you think so ?
> Hence my problem with wham using absolute coordinate instead of actual distance between the two centers of mass
> What do you suggest to retrieve the actual values of my reaction coordinate without rerrunning everything ?

It should be straightforward to apply a systematic shift to the values 
in the output PMF curve. But I don't know how you've set up your pull 
code to get such output in the first place. The absolute position of 
group0 should be totally irrelevant.

-Justin

> Serge
>
>
> Dear Justin
>
> Thanks for your fast answer
> OK for the larger group0 COM but as you say my reaction coordinate is the distance between the two COMs - The problem being that in the profile.xvg file the X coordinate reflects the box sizes and thus I can't superimpose the profiles coming from Simulations 1) and 2) - I need to have in the pmf profile X coordinates equal to the distance between COMs not absolute distance as it is now .. How can I correct the x coordinate in the pmfs so that they represent my real reaction coordinate which is the distance between COMs?
> Isn't it a strange choice to write the absolute coordinates of the COM in the pullx files instead of the real reaction coordinate we all need ?  Now do I have to reread all my trajs, calculate the center of mass of the X coordinate of the COM of DNA (my fixed molecule) and subtract it at each line of the pullx files ... This does not make sense ...
>
> Thanks a lot
>
>
> On 10/31/18 11:43 AM, CROUZY Serge 119222 wrote:
>> Dear gromacs users
>>
>> I've been running gromacs for several years, and enjoyed it..
>> I've been running PMF calculations for protein DNA interactions without problems until recently:
>> Now I'm puzzled with the X coordinates being written in the pullx
>> files (and thus taken as reaction coordinate values in the PMF). This
>> is my problem
>>
>> I run two simulations 1)  protein A moving away from DNA                      DNA---A->  along x
>>                                          2) same protein A moving away
>> from DNA in the presence of protein B:  B-DNA---A -> along x In 2) A
>> and B interact slightly and I expect to see a slight difference in the
>> profiles for pulling A away In both simulations I'm pulling on center
>> of mass of A away from center of mass of DNA (force along X only)
>>
>> My problem is that the x pulling coordinates in simulation 2) (in the pullx files) are around 9 A larger than in simulation 1). Consequence: the profiles are shifted along x by around 9 A. This is not logical to me since my reaction coordinate is distance between com of A and com of DNA which should be the same !!
>> It's as if the size of the simulation box (around 10 A larger in 2) to
>> accommodate B ) mattered .. ?!  (I'm running PME with PBC in water...)
> Yes, that makes sense. The first column(s) in pullx.xvg are whatever
> (x,y,z) components of the group0 COM. If your box is larger, then naturally the position is different. What should not be different is the relative DNA-A distance, which is the actual reaction coordinate.
> Absolute coordinates of any one species don't really matter.
>
> -Justin
>
> ************************************
>
>
> ------------------------------
>
> Message: 5
> Date: Wed, 31 Oct 2018 13:24:45 -0400
> From: Justin Lemkul <jalemkul at vt.edu>
> To: gmx-users at gromacs.org
> Subject: Re: [gmx-users] 3. Re: Strange pullx coordinates (PMF
> 	calculations) (Justin Lemkul)
> Message-ID: <10376386-bbb5-fa99-9743-8786e006974c at vt.edu>
> Content-Type: text/plain; charset=utf-8; format=flowed
>
>
>
> On 10/31/18 12:54 PM, CROUZY Serge 119222 wrote:
>> Dear Justin
>>
>> Thanks for your fast answer
>> OK for the larger group0 COM but as you say my reaction coordinate is
>> the distance between the two COMs - The problem being that in the profile.xvg file the X coordinate reflects the box sizes and thus I can't superimpose the profiles coming from Simulations 1) and 2) - I need to have in the pmf profile X coordinates equal to the distance between COMs not absolute distance as it is now .. How can I correct the x coordinate in the pmfs so that they represent my real reaction coordinate which is the distance between COMs?
>> Isn't it a strange choice to write the absolute coordinates of the COM in the pullx files instead of the real reaction coordinate we all need ?  Now do I have to reread all my trajs, calculate the center of mass of the X coordinate of the COM of DNA (my fixed molecule) and subtract it at each line of the pullx files ... This does not make sense ...
> There are two columns in pullx.xvg - X and dX, the reference position and the distance along the biasing potential. WHAM should be using the latter; the absolute position in space should not matter. You can also use pullf.xvg files to construct the PMF. The reference values for each window should be taken from each .tpr file.
>
> -Justin
>
>> Thanks a lot
>>
>>
>> On 10/31/18 11:43 AM, CROUZY Serge 119222 wrote:
>>> Dear gromacs users
>>>
>>> I've been running gromacs for several years, and enjoyed it..
>>> I've been running PMF calculations for protein DNA interactions without problems until recently:
>>> Now I'm puzzled with the X coordinates being written in the pullx
>>> files (and thus taken as reaction coordinate values in the PMF). This
>>> is my problem
>>>
>>> I run two simulations 1)  protein A moving away from DNA                      DNA---A->  along x
>>>                                           2) same protein A moving
>>> away from DNA in the presence of protein B:  B-DNA---A -> along x In
>>> 2) A and B interact slightly and I expect to see a slight difference
>>> in the profiles for pulling A away In both simulations I'm pulling on
>>> center of mass of A away from center of mass of DNA (force along X
>>> only)
>>>
>>> My problem is that the x pulling coordinates in simulation 2) (in the pullx files) are around 9 A larger than in simulation 1). Consequence: the profiles are shifted along x by around 9 A. This is not logical to me since my reaction coordinate is distance between com of A and com of DNA which should be the same !!
>>> It's as if the size of the simulation box (around 10 A larger in 2)
>>> to accommodate B ) mattered .. ?!  (I'm running PME with PBC in
>>> water...)
>> Yes, that makes sense. The first column(s) in pullx.xvg are whatever
>> (x,y,z) components of the group0 COM. If your box is larger, then naturally the position is different. What should not be different is the relative DNA-A distance, which is the actual reaction coordinate.
>> Absolute coordinates of any one species don't really matter.
>>
>> -Justin
>>

-- 
==================================================

Justin A. Lemkul, Ph.D.
Assistant Professor
Virginia Tech Department of Biochemistry

303 Engel Hall
340 West Campus Dr.
Blacksburg, VA 24061

jalemkul at vt.edu | (540) 231-3129
http://www.thelemkullab.com

==================================================