[gmx-users] Re: rotational correlation function

[David van der Spoel]

On 2010-09-23 20.38, Paymon Pirzadeh wrote:
> Great tips and advices. I appreciate your attention.
> But, what would the g_rms -fit rot+trans  do?

compute the rmsd. something different. I said trjconv.
>
> Paymon
>
>
>
> On Thu, 2010-09-23 at 20:07 +0200, David van der Spoel wrote:
>> On 2010-09-23 18.14, Paymon Pirzadeh wrote:
>>> Dear Dr. van der Spoel,
>>> Thanks a lot for your clarification. But some questions yet:
>>>
>>> 1. If I do not provide the "-aver" option, will the program calculate
>>> the C(t) for individual N-H bond?
>>>
>> Yes.
>>
>>> 2. In a paper that I just read, JCP vol. 131 155103 (2009),  they claim
>>> they have calculated the Ctot(t) which could be factorized into
>>> rotational and internal components, Ctot(t)=Crot(t)*Cint(t). I assume
>>> g_rotacf gives the rotational part. How one can calculate the Ctot(t)
>>> (they claim they calculatetd the Ctot and Crot with GROMACS utilities)?
>>> How can we generate the rotation matrix (omega(t)) and use it? What's
>>> the use of "-fitfn"? Is it sth that could be done in other data
>>> processing softwares?
>> If your sim is long enough (10-20 times the rotation time, which for a
>> small protein is on the order of a couple of ns) you can fit all the
>> individual NH -acfs to two exponentials, 1 for the NH relaxation and one
>> for the overall.
>>
>>>
>>> 3. How can we generate a trajectory with no internal motions? Does it
>>> have anything to do with g_rms?
>> You can do the reverse by fitting trjconv -fit rot+trans this will give
>> you a protein with no overall rotation+translation from which you can
>> derive the order parameters and relaxation times of individual NH vectors.
>>
>>
>> Check this ref too:
>> @Article{Villa2006a,
>>      author  = {A. Villa and G. Stock},
>>     title   = {What NMR relaxation can tell us about the internal motion
>> of an RNA
>>    hairpin: A molecular dynamics simulation study},
>>     journal = {J. Chem. Theory Comput.},
>>     volume  = {2},
>>     year    = {2006},
>>     pages   = {1228-1236},
>> }
>>
>>>
>>> Sorry for mass of questions.
>>> Regards,
>>>
>>> Paymon
>>>
>>>
>>> On Thu, 2010-09-23 at 08:20 +0200, David van der Spoel wrote:
>>>> On 2010-09-23 01.27, Paymon Pirzadeh wrote:
>>>>> Dear Dr. Chaban,
>>>>> I meant the N-H bond vectors of protein backbone for calculation of
>>>>> rotational time correlation function to calculate the rotational
>>>>> diffusion constant of my protein. I need a protocol which walks me step
>>>>> by step through the procedure.
>>>>> Thanks for your Attention.
>>>> define all NH bonds in an index file as
>>>> [ NH ]
>>>> 5 6
>>>> 10 11
>>>> 23 24
>>>>
>>>> etc.
>>>>
>>>> Then run g_rotacf -P 2 [ more flags ] -aver
>>>> Analyze the ACF to extract the order parameters.
>>>> You will need to read Lipari&   Szabo papers:
>>>> @Article{ Lipari82a,
>>>>      author =       "G. Lipari and A. Szabo and R. M. Levy",
>>>>      title =        "Protein Dynamics and {NMR} relaxation: comparison of
>>>>                      simulations with experiment",
>>>>      journal =      BTnature,
>>>>      year =         1982,
>>>>      volume =       300,
>>>>      pages =        "197-198"
>>>> }
>>>>
>>>> @Article{ Lipari82b,
>>>>      author =       "G. Lipari and A. Szabo",
>>>>      title =        "Model-Free Approach to the Interpretation of Nuclear
>>>>                      Magnetic Resonance Relaxation in
>>>>                      Macromolecules. 1. Theory and Range of Validity",
>>>>      journal =      BTjacs,
>>>>      year =         1982,
>>>>      volume =       104,
>>>>      pages =        "4546-4559"
>>>> }
>>>>
>>>>
>>>>>
>>>>> Paymon
>>>>>
>>>>>
>>>>> On Wed, 2010-09-22 at 19:03 -0400, Vitaly Chaban wrote:
>>>>>>> Message: 2
>>>>>>> Date: Wed, 22 Sep 2010 12:07:27 -0600
>>>>>>> From: Paymon Pirzadeh<ppirzade at ucalgary.ca>
>>>>>>> Subject: [gmx-users] rotational correlation function
>>>>>>> To: gmx-users at gromacs.org
>>>>>>> Message-ID:<1285178847.11669.66.camel at paymon-desktop>
>>>>>>> Content-Type: text/plain
>>>>>>>
>>>>>>> Hello,
>>>>>>> Imagine if the total correlation function of a protein could be
>>>>>>> factorized into rotational and internal portions:
>>>>>>> Ctot(t)=Crot(t)*Cint(t). With which commands I can calculate the Ctot(t)
>>>>>>> and with which I can get the Crot(t)? Does g_rotacf produce the Crot(t)?
>>>>>>> What does the g_rms -fit do? I am confused on how I can use gromacs
>>>>>>> commands on my output files. Sorry for mass of e-mails.
>>>>>>>
>>>>>>> Paymon
>>>>>>
>>>>>>
>>>>>> Hey, Paymon -
>>>>>>
>>>>>> Do you mean the correlation function of the velocity (VACF)? Why not
>>>>>> to use g_velacc and g_rotacf for translational and rotational VACFs? I
>>>>>> suppose, your molecule is rigid (enough) for rotational velocoty to
>>>>>> make sense.
>>>>>>
>>>>>> --
>>>>>> Dr. Vitaly V. Chaban
>>>>>> Department of Chemistry
>>>>>> University of Rochester
>>>>>> Rochester, NY 14627-0216
>>>>>> United States of America
>>>>>
>>>>
>>>>
>>>> --
>>>> David van der Spoel, Ph.D., Professor of Biology
>>>> Dept. of Cell&   Molec. Biol., Uppsala University.
>>>> Box 596, 75124 Uppsala, Sweden. Phone:	+46184714205.
>>>> spoel at xray.bmc.uu.se    http://folding.bmc.uu.se
>>>
>>
>>
>> --
>> David van der Spoel, Ph.D., Professor of Biology
>> Dept. of Cell&  Molec. Biol., Uppsala University.
>> Box 596, 75124 Uppsala, Sweden. Phone:	+46184714205.
>> spoel at xray.bmc.uu.se    http://folding.bmc.uu.se
>


-- 
David van der Spoel, Ph.D., Professor of Biology
Dept. of Cell & Molec. Biol., Uppsala University.
Box 596, 75124 Uppsala, Sweden. Phone:	+46184714205.
spoel at xray.bmc.uu.se    http://folding.bmc.uu.se