[gmx-users] pbc atom

[Justin A. Lemkul]

Gavin Melaugh wrote:
> There are 168 atom types in my system, 2 molecules comprising of 84
> atoms each. The average temperature is 600 K but there are large
> fluctuations ca ~100 K, which I assume is expected with such a smal
> system. Ill check the energies
> 

I would think it possible, even likely, that the thermostat is driving the 
behavior of your system, since it contains so few atoms.  Do check to see if 
temperature spikes correlate with movement of your pull_group to one side of the 
sampling window.  If it is the problem, try a different thermostat (or at least 
longer tau_t with N-H), and/or a substantially stronger pull_k1 if you're sure 
you need/want N-H as your thermostat.

I certainly agree with Chris' proposed test with two water molecules.  If you 
see the same thing, I'm even more confident that I'm right :)

-Justin

> Gavin
> 
> Justin A. Lemkul wrote:
>>
>> Gavin Melaugh wrote:
>>> O.K Thanks
>>>
>>> My apologies for not stating earlier that I had the box dimensions set
>>> to zero but I thought that when you set pbc = no and ns_type = simple
>>> that "ignore the box" (from manual) was implemented.
>>>
>> Indeed.
>>
>> How many atoms are in your system?  How well-conserved is the
>> temperature?  I imagine that if you only have a few atoms (relative to
>> a normal condensed-phase system), the species involved would be fairly
>> strongly impacted by the (potentially) wide oscillations of the
>> Nose-Hoover thermostat.  Maybe the thermostat is kicking your system
>> around?  Check the relevant energies and see if they sync up with the
>> configurations moving back and forth.
>>
>> -Justin
>>
>>> Gavin
>>>
>>> Justin A. Lemkul wrote:
>>>> Gavin Melaugh wrote:
>>>>> Justin
>>>>>
>>>>> In regard to my query regarding the cut-offs at zero with no pbc. I
>>>>> assume its is fine to have box dimensions (0 0 0) to comply with the
>>>>> definition of a vacuum simulation. I am just worried that there may be
>>>>> some default cut-off that I am not aware of.
>>>>>
>>>> You have a box of zero size?  That doesn't make any sense to me, and I
>>>> doubt that such a setup plays nice with the MD code.  Frankly, I don't
>>>> see how it worked at all.
>>>>
>>>> What if you place your system in a big box, no PBC, cutoffs = 0,
>>>> etc? Does that give you a sensible result?
>>>>
>>>>> Cheers
>>>>>
>>>>> P.S I have plotted the pullx.xvg files for the distances between the
>>>>> COMs of the molecules in the problematic windows. As was the case with
>>>>> plotting with g_dist there is significant population of the at the
>>>>> extremities of the window compared to that around the average r0.
>>>>>
>>>> Then technically, the average is probably fine, so the pull code is
>>>> satisfied. The sampling is clearly messed up for some reason.  Keep
>>>> digging :)
>>>>
>>>> -Justin
>>>>
>>>>> Justin A. Lemkul wrote:
>>>>>> chris.neale at utoronto.ca wrote:
>>>>>>> I have seen this before, and there are a few possible reasons, but
>>>>>>> I'm still waiting to see that histogram with two peaks
>>>>>>> (http://lists.gromacs.org/pipermail/gmx-users/2010-August/053311.html).
>>>>>>>
>>>>>>>
>>>>>>> It's really hard to help you when we have to guess what the problem
>>>>>>> looks like... or perhaps you posted it and I missed it?
>>>>>> The histogram was posted several days ago:
>>>>>>
>>>>>> http://lists.gromacs.org/pipermail/gmx-users/2010-August/053317.html
>>>>>>
>>>>>> The attachment link is:
>>>>>>
>>>>>> http://lists.gromacs.org/pipermail/gmx-users/attachments/20100820/fc07fb81/histo.ps.bin
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>> -Justin
>>>>>>
>>>>>>> -- original message --
>>>>>>>
>>>>>>> O.K thanks anyway
>>>>>>>
>>>>>>>
>>>>>>> Justin A. Lemkul
>>>>>>> wrote:http://lists.gromacs.org/pipermail/gmx-users/2010-August/053311.html
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>> Gavin Melaugh wrote:
>>>>>>>>> Thanks Justin
>>>>>>>>>
>>>>>>>>> Have you any idea why when generating umbrella histograms for the
>>>>>>>>> pmf I
>>>>>>>>> would get two peaks in the histograms above a distance of 2 nm,
>>>>>>>>> but
>>>>>>>>> below 2 nm I get well behaved histograms that lead to a very
>>>>>>>>> good profile in the pmf. To the best of my knowledge the
>>>>>>>>> configurations
>>>>>>>>> are all very well equilibrated at their respective COM distances.
>>>>>>>>> Umbrella sampling is performed on all windows using a force
>>>>>>>>> constant of
>>>>>>>>> 1000 kj/mol at 600 K.
>>>>>>>>>
>>>>>>>> Sorry, no clue.
>>>>>>>>
>>>
> 
> 

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

Justin A. Lemkul
Ph.D. Candidate
ICTAS Doctoral Scholar
MILES-IGERT Trainee
Department of Biochemistry
Virginia Tech
Blacksburg, VA
jalemkul[at]vt.edu | (540) 231-9080
http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin

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