[gmx-users] g_hbond and g_rdf in vacuum

[Justin Lemkul]

On 10/28/13 3:30 AM, Santu Biswas wrote:
>>> "Not working" is too vague a symptom for anyone to guess what the
> problem
>>> is, sorry.
>>>
>>> Mark
>>> On Oct 24, 2013 9:39 AM, "Santu Biswas" <santu.biswas85 at gmail.com>
>> wrote:
>>>
>>>> dear users,
>>>>
>>>>                I am performing 500ps mdrun in vacuum for
>>> polypeptide(formed
>>>> by 10-residues leucine) using gromacs_4.5.5(double-
> precision) using
>>>> opls-aa/L force field.Input file for 500ps mdrun is given below
>>>>
>>>>
>>>> title                    = peptide in vaccum
>>>> cpp                    = /lib/cpp
>>>>
>>>> ; RUN CONTROL
>>>> integrator             = md
>>>> comm_mode        = ANGULAR
>>>> nsteps                 = 500000
>>>> dt                        = 0.001
>>>> ; NEIGHBOR SEARCHING
>>>> nstlist                  = 0
>>>> ns_type               = simple
>>>> pbc                     = no
>>>> rlist                     = 0
>>>> ; OUTPUT CONTROL
>>>> nstxout                  = 1000
>>>> nstvout                  = 1000
>>>> nstxtcout               = 0
>>>> nstlog                    = 1000
>>>> constraints             = none
>>>> nstenergy               = 1000
>>>> ; OPTION FOR ELECTROSTATIC AND VDW
>>>> rcoulomb                 = 0
>>>> ; Method for doing Van der Waals
>>>> rvdw                        = 0
>>>> ; OPTIONS FOR WEAK COUPLING ALGORITHMS
>>>> tcoupl                      = V-rescale
>>>> tc_grps                    = Protein
>>>> tau_t                        = 0.1
>>>> ref_t                         = 300
>>>> gen_vel                    = yes
>>>> gen_temp                 = 300
>>>>
>>>> Using the 500ps trajectory if i run g_hbond_d for calculating the
>> number
>>> of
>>>> hydrogen bonds as a function of time using index file(where atom O and
>>> atom
>>>> N H is used) it is not working.
>>>> Also if i used g_rdf_d with pbc=no using the 500ps trajectory it is
>> also
>>>> not working.
>>>> I do not know why this is happening.
>>>>
>>>> ------
>>>> santu
>>>> --
>>> Thanks Mark for your reply.
>>>
>>
>>     Using the 500ps trajectory i want to calculate the number of hydrogen
>> bonds as a function of time in vacuum .For this calculation i
>>
>>>   have uesd
>>> g_hbond_d  -f  traj_0-500ps.trr  -s 500ps.tpr -n index.ndx -num
>>> hbond-num.xvg -dist dist.xvg -ang angle.xvg
>>
>>
> With what groups? Can there be any hydrogen bonds between those groups?
>
> Is there a bug fixed in a version of g_hbond that isn't 2 years old? Did a
> shorter trajectory work because it took less time? Does doing only one of
> three analyses help things to work? You'd be much closer to a solution if
> you'd tried some simplifications and done some detective work already ;-)
>
>
>>> Programm was running .After 1 hour it was still running but there was no
>>> output.
>>>
>>
>>      If I calculate the number of hydrogen bonds as a function of time in
>> water (no vacuum) using the same command line then there was  no problem.
>>
>>     Same problem when I used g_rdf in vacuum.The commad line I have used
>>     g_rdf_4.5.5 -f traj.trr -s 500ps.tpr -n index.ndx -o rdf.xvg and also
>> checked with -nopbc with the same command line.
>>
>
>
> RDF of what, in vacuum? What groups did you use?
>
>
>>     The programm is running but in the output file nothing is written.
>>     If I used g_rdf in water using the same command line there was no
>> problem.
>>
>
> OK - but does your analysis make sense in vacuum?
>
> Mark
>
> Thanks Mark for your reply
>
>
> In the index file I mentioned C=O oxygen atom and N-H hydrogen atom. Now I
> want to calculate the distance between O atom and N-H hydrogen atom using
> g_rdf in vacuum.
>
>
> Similarly using the same index file I want to calculate the number of
> hydrogen bonds as a function of time in vacuum using g_hbond.
> Yes there is a H bond between C=O oxygen atom of i-th residue and N-H
> hydrogen of i+4-th residue.
>   I also checked with shorter trajectory there is also same problem .
>
> Please let me know If there are any mistakes.
>

Does your system have a non-zero size?  That is, are all the box vectors a 
somewhat reasonable value?  They are, of course, irrelevant for the simulation 
(assuming pbc = no for in vacuo), but for analysis, you need a sensible box.  If 
you don't, you can recenter the trajectory using trjconv -box -center.

-Justin

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

Justin A. Lemkul, Ph.D.
Postdoctoral Fellow

Department of Pharmaceutical Sciences
School of Pharmacy
Health Sciences Facility II, Room 601
University of Maryland, Baltimore
20 Penn St.
Baltimore, MD 21201

jalemkul at outerbanks.umaryland.edu | (410) 706-7441

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