Thanks Jason, I use amber10, it should work. To begin, I will try to
run sander several times manually with ig= -1 and compare the final
structures. Then may be, a good procedure to avoid to have the same seed
should be to use a loop in the refinement script to return to the first
line after a simulation (or to use an external script). For the moment,
I don't find any information about how NMR modellers classically launch
their runs with hundreds of structures e.g.
Best wishes,
Baptiste
Le 07/09/2010 17:31, Jason Swails a écrit :
> Setting ig=-1 in certain versions of amber (at least 10 and later) will pull
> the random seed from the system clock, which is the closest you can get. If
> your system clock has coarse precision and you plan to submit all of your
> jobs at once using a script, some of your simulations may end up with the
> same seed, so you may want to investigate a "pause" or "wait" command in
> whatever language you're using.
>
> Good luck!
> Jason
>
> On Tue, Sep 7, 2010 at 10:09 AM, Baptiste Legrand<bap.legrand.gmail.com>wrote:
>
>> Thank you Jason and David for your replies. Is it possible to
>> automatically generate random ig values and restart a new refinement?
>> Thanks
>>
>> Best
>> Baptiste
>>
>>
>> Le 07/09/2010 14:03, case a écrit :
>>> On Tue, Sep 07, 2010, Baptiste Legrand wrote:
>>>> I'm a new amber user (I'm more familiar with cns). At the moment, I'm
>>>> following the tutorial4 on the nmr refinement on a DNA duplex to
>>>> determine a modified peptide conformation. All is fine but I'm quite
>>>> surprised to have only one structure at the end of the sander run and
>>>> not a set of structure. Is there an option to activate or a more
>>>> specific protocol to calculate several structures?
>>> In Amber, you are expected to run the refinement many times (with random
>>> initial velocites, or different starting structures) in order to obtain a
>>> family of structures.
>>>
>>> ....dac
>>>
>>>
>>> _______________________________________________
>>> AMBER mailing list
>>> AMBER.ambermd.org
>>> http://lists.ambermd.org/mailman/listinfo/amber
>>
>> _______________________________________________
>> AMBER mailing list
>> AMBER.ambermd.org
>> http://lists.ambermd.org/mailman/listinfo/amber
>>
>
>
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Tue Sep 07 2010 - 10:00:03 PDT
