[gmx-users] pca-based MD

[James Starlight]

Dear all!

I've read some reference papers about EDA sampling methods and found
such usefull things. First of all as  I understood for biassing MD
simulation along several PCs extracted from another run the make_edi
-radacc 1-3   option is exactly what I need.

But I havent still understood about missmatching of atom numbers
between EDA (done on C-alpha atoms)  and biassed_MD ( all atoms of
system including solvent).

E.g whan I
ve run mdrun

I've obtain error

Fatal error:
Nr of atoms in pca_biased.edi (3075) does not match nr of md atoms (46185)

where 3075 is n atoms of my protein (althought I've extracted
eigenvectors from c-alpha only) and 46185- full system includding
membrane and SOL in that case.

Does it mean that such method is only applicable for biassed
simulations of the systems  in vacuu or with implicit solvent ?

James

2012/9/24 James Starlight <jmsstarlight at gmail.com>:
> I've still made such 'only c-alpha ensemble' of my structures by the
> other software and performed x-ray PCA. As the result I've extracted
> eigenvectors and obtained reasonable distribution (projection) of the
> structures along that eigenvectors.
>
> Now I have questions about pca-biassed MD_run. I've made *.edi file
> from eigenvectors calculated based on the c-alpha atoms of my x-ray
> ensemble. How I could use it with my system consisted of much more
> atoms (full atomic protein + solvent) than smaller c-alpha subset of
> the x-ray data (only c-alpha atoms from the same protein) ?
> i.e if I run mdrun -v -deffnm MD -ei sam.edi
>
> I obtain error about mismatching of atom number from the edi as well
> as system.tpr .
> Is there any way to extrapolate number of atoms in the sam.edi ?
>
> James
>
> 2012/9/23, James Starlight <jmsstarlight at gmail.com>:
>> I've tried to make PCA from my X-ray data and forced with many problems :)
>>
>> Firstly I've made pdb trajectory in NMR-like format ( by means of
>> pymol) consisted of all X-ray structures.
>>
>> than I've make .tpr file (From the tpr of the same protein which I've
>> simulated previously) for the subset of C-alpha atoms common to all
>> structures
>>
>> Finally I've tried to calculate eigenvectors
>> Structure or trajectory file has more atoms (2196) than the topology (302)
>>
>> Does it mean that all structures in trajectory must have only C-alpha
>> atoms initialy ?
>>
>> IS there another way to make tpr as well as pdb trajectory files for
>> such x-ray PCA?
>>
>>
>> James
>>
>> 2012/9/24 Thomas Evangelidis <tevang3 at gmail.com>:
>>> Hi,
>>>
>>> thanks again for explanation. Its also intresting to me is it possible
>>>> to do further biassed MD guided on that FMA modes as well as obtain
>>>> projections onto that FMA sub-spaces of X-ray datasets for instance ?
>>>> ( e.g for comparison of the results from FMA of experimental data as
>>>> well as MD_data)
>>>>
>>>>
>>>  Have a look at another thread posted today, named "PC comparison between
>>> two simulations".
>>>
>>> On a second thought, you might want to consider the nudged elastic band
>>> method and its variants for your case, since you have the initial, the
>>> final and intermediate states of your protein. Unfortunately they are not
>>> implemented in GROMACS, but they are in AMBER.
>>>
>>> Thomas
>>>
>>> --
>>>
>>> ======================================================================
>>>
>>> Thomas Evangelidis
>>>
>>> PhD student
>>> University of Athens
>>> Faculty of Pharmacy
>>> Department of Pharmaceutical Chemistry
>>> Panepistimioupoli-Zografou
>>> 157 71 Athens
>>> GREECE
>>>
>>> email: tevang at pharm.uoa.gr
>>>
>>>           tevang3 at gmail.com
>>>
>>>
>>> website: https://sites.google.com/site/thomasevangelidishomepage/
>>> --
>>> gmx-users mailing list    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.
>>> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>>