[gmx-users] Re: Re: g_sham problem

[David van der Spoel]

On 2012-07-08 15:44, Justin A. Lemkul wrote:
>
>
> On 7/8/12 9:39 AM, siddhant jain wrote:
>> Thanks Justin
>> Also I wanted to ask that to know my structure which has minimum
>> energy, can
>> I take the corresponding values of principle components, look at the
>> time at
>> which they occur simultaneously, and hence note the structure
>> corresponding
>> to that time. Will it be reasonable enough or its vague.
>>
>
> You can certainly identify structures within different regions of the
> free energy surface in this way.  Just be careful about saying any
> structure has a particular "energy" value, since this can mean a number
> of different things.
>
> -Justin
>
Indeed, in particular since eigenvectors from PCA do not cover the whole 
space of motions. If you take other coordinates, e.g. the RMSD and the 
radius of gyration, where you do cover all possibilities in a 2D plot 
this can work. Note however that this is complicated too, because the 
volume of the bins in your 2D plot are not the same in general, so even 
if you have more points in one bin then in another it might be because 
the volume is larger, rather than that it is more populated.

-- 
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