[gmx-users] using more processors for g_mindist

Mark Abraham Mark.Abraham at anu.edu.au
Wed Feb 29 13:02:48 CET 2012


On 29/02/2012 9:05 PM, aiswarya pawar wrote:
> Mark,
>
> In the g_mindist output data gives which atom was within cutoff to 
> protein atom ie the output file of g_mindist

What makes you think there is only one atom satisfying that cutoff?

>
> 5.000000e+02           518         42680
> 5.010000e+02           518         20942
> 5.020000e+02           518         67844
> 5.030000e+02           518          5984
> 5.040000e+02           518         67844
> 5.050000e+02           518          5984
> 5.060000e+02           518         30116
> 5.070000e+02           518         67844
> 5.080000e+02           518         32957
> 5.090000e+02           518         67844
> 5.100000e+02           518         19610
> 5.110000e+02           518         19610
> 5.120000e+02           518         22895
> 5.130000e+02           518         30116
> 5.140000e+02           518         19610
> 5.150000e+02           518         22895
> 5.160000e+02           518         13628
> 5.170000e+02           518          5984
>
> 2nd column is the protein atom and the 3rd column the water atom.

Chunk of output from an unnamed file without giving the command line 
that generated it is close to useless :-) Anyway, whatever information 
is present here is not helping you observe the duration of a contact 
between the protein and any given water molecule.

>
> if iam doing g_dist how is it possible to know the which water 
> molecule has to be considered for the distance calculations ie should 
> i compute the distance between one protein atom against all the water 
> within the cut off individually?

If you have read g_dist -h and want to ask this question, then I did not 
understand your previous statement of your objective.

Mark

>
> Thanks
>
> On Wed, Feb 29, 2012 at 2:12 PM, Mark Abraham <Mark.Abraham at anu.edu.au 
> <mailto:Mark.Abraham at anu.edu.au>> wrote:
>
>     On 29/02/2012 7:21 PM, aiswarya pawar wrote:
>>     Mark,
>>
>>     i want to know which water atoms stay within a cut off to protein
>>     atom.
>
>     OK, but as you will read in g_mindist -h, it will count such atoms
>     and not identify which they were...
>
>
>>     ie i need the duration at which a water resides on the protein atoms.
>
>     ... and identity of the atoms is needed for measuring duration of
>     contact. g_dist does something like this, and reading g_mindist -h
>     should have prompted you to find this out.
>
>     Otherwise, you will have to construct an index group for each
>     water molecule, and script a loop to examine each water molecule
>     separately using some tool that observes what you really want to
>     measure.
>
>
>>     so for that i need the whole 5ns frames because am looking for
>>     water molecules which reside more than 50% of time.
>
>     You also need to be clear about whether you care about continuous
>     contact. Does a water molecule that oscillates at a distance
>     around the cutoff reside about 50% of the time?
>
>     You still don't need high time resolution for testing whether this
>     analysis might give you the information you want. The water
>     molecules that are in contact for more than 50% of the time
>     (continuous or not) will show up in 5 snapsnots spaced every
>     nanosecond. 5000 snapshots every picosecond is better, but not if
>     you can't afford to wait for it.
>
>
>     Mark
>
>>
>>     Thanks
>>
>>     On Wed, Feb 29, 2012 at 12:41 PM, Mark Abraham
>>     <Mark.Abraham at anu.edu.au <mailto:Mark.Abraham at anu.edu.au>> wrote:
>>
>>         On 29/02/2012 6:01 PM, aiswarya pawar wrote:
>>>         Mark,
>>>
>>>         Right now am computing distance between each protein atom
>>>         against all water atoms,
>>
>>         That's expensive. mdrun goes to great lengths to speed up
>>         computing billions of distances.
>>
>>>         which is taking too long for 5ns run. i cant reduce the frames
>>
>>         Yes you can. Even if you think you need data from every
>>         frame, you probably don't because they're correlated with
>>         each other, and at the very least you can do a pilot study on
>>         a frame every 100ps or every nanosecond before committing to
>>         one on all the frames.
>>
>>
>>>         either the number of water atoms. So is there any alternate.
>>>
>>
>>         You are not likely to get a better solution if you only
>>         describe your attempt, rather than describe the objective.
>>         Asking "how do I hammer harder?" if you're hammering a screw
>>         makes it impossible to get the correct solution "Use a
>>         screwdriver".
>>
>>         Mark
>>
>>
>>>         Thanks
>>>
>>>         On Wed, Feb 29, 2012 at 12:27 PM, Mark Abraham
>>>         <Mark.Abraham at anu.edu.au <mailto:Mark.Abraham at anu.edu.au>>
>>>         wrote:
>>>
>>>             On 29/02/2012 5:17 PM, aiswarya pawar wrote:
>>>>             Dear all,
>>>>
>>>>             Am running g_mindist on large number of atoms, i would
>>>>             like to know whether i can run this on more than one
>>>>             processors say 8 processors to speed up the task?
>>>
>>>             No. If it will take too long, you need to reduce your
>>>             number of frames (trjconv), or the number of atoms (also
>>>             trjconv), or some such.
>>>
>>>             Mark
>>>
>>>
>>>>             and will this effect the output in anyways.
>>>>
>>>>             Thanks,
>>>>
>>>>             -- 
>>>>             Aiswarya  B Pawar
>>>>
>>>>
>>>>
>>>>
>>>
>>>
>>>             --
>>>             gmx-users mailing list gmx-users at gromacs.org
>>>             <mailto:gmx-users at gromacs.org>
>>>             http://lists.gromacs.org/mailman/listinfo/gmx-users
>>>             Please search the archive at
>>>             http://www.gromacs.org/Support/Mailing_Lists/Search
>>>             before posting!
>>>             Please don't post (un)subscribe requests to the list.
>>>             Use the
>>>             www interface or send it to
>>>             gmx-users-request at gromacs.org
>>>             <mailto:gmx-users-request at gromacs.org>.
>>>             Can't post? Read
>>>             http://www.gromacs.org/Support/Mailing_Lists
>>>
>>>
>>>
>>>
>>>         -- 
>>>         Aiswarya  B Pawar
>>>
>>>         Project Assistant,
>>>         Bioinformatics Dept,
>>>         Indian Institute of Science
>>>         Bangalore
>>>
>>>
>>>
>>>
>>
>>
>>         --
>>         gmx-users mailing list gmx-users at gromacs.org
>>         <mailto:gmx-users at gromacs.org>
>>         http://lists.gromacs.org/mailman/listinfo/gmx-users
>>         Please search the archive at
>>         http://www.gromacs.org/Support/Mailing_Lists/Search before
>>         posting!
>>         Please don't post (un)subscribe requests to the list. Use the
>>         www interface or send it to gmx-users-request at gromacs.org
>>         <mailto:gmx-users-request at gromacs.org>.
>>         Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>>
>>
>>
>>
>>     -- 
>>     Aiswarya  B Pawar
>>
>>     Project Assistant,
>>     Bioinformatics Dept,
>>     Indian Institute of Science
>>     Bangalore
>>
>>
>>
>>
>
>
>     --
>     gmx-users mailing list gmx-users at gromacs.org
>     <mailto:gmx-users at gromacs.org>
>     http://lists.gromacs.org/mailman/listinfo/gmx-users
>     Please search the archive at
>     http://www.gromacs.org/Support/Mailing_Lists/Search before posting!
>     Please don't post (un)subscribe requests to the list. Use the
>     www interface or send it to gmx-users-request at gromacs.org
>     <mailto:gmx-users-request at gromacs.org>.
>     Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>
>
>
>
> -- 
> Aiswarya  B Pawar
>
> Project Assistant,
> Bioinformatics Dept,
> Indian Institute of Science
> Bangalore
>
>
>
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-users/attachments/20120229/e508ab52/attachment.html>


More information about the gromacs.org_gmx-users mailing list