Re: [AMBER] A question about PCA analysis

From: Daniel Roe <daniel.r.roe.gmail.com>
Date: Tue, 25 Oct 2016 08:28:16 -0400

On Tue, Oct 18, 2016 at 6:06 AM, Karolina MitusiƄska (Markowska)
<markowska.kar.gmail.com> wrote:
> 1. What can I learn from the histograms? What is on the x and y axis?

As used in the tutorials the histograms are being used to see whether
the system is undergoing the same kinds of motions to the same degree
in each trajectory. See e.g.
http://pubs.acs.org/doi/abs/10.1021/ct400862k and/or
http://pubs.acs.org/doi/abs/10.1021/jp4125099 for more examples of
this kind of analysis with cpptraj.

> 2. Is 'createcrd' command necessary? If I am using just one trajectory file
> do I need to 'createcrd' and everytime use 'crdaction'? For my short, test
> simulation I got a message from cptraj:
> 1ryi-trajectories "1ryi-trajectories" (coordinates), size is 10000
> (5361.33 MB) Box Coords, 46846 atoms

It's not necessary, but it can be more convenient. As Hai mentioned,
it is often faster to re-iterate over a trajectory once it has been
loaded in to memory. Also, when you project coordinates onto
eigenvectors you need to make sure they are the same as the
coordinates used to create the covariance matrix that you obtained the
eigenvectors from. If you save the coordinates into memory after e.g.
RMS-fitting you don't need to RMS-fit again, whereas if you read the
coordinates from the disk again you do.

> 3. Are the frames obtained after projection real frames from my simulation?

Not sure what you mean here - you'll have to clarify.

> 4. A friend of mine found this paper:
> http://onlinelibrary.wiley.com/doi/10.1002/chem.201602923/abstract and I'm
> curious, how could I calculate the vibrational modes in my simulations?
> Which command should I use for that?

You probably want to use the 'nmode' function from 'nab' (see the
Amber 16 manual, section 40.4). Be aware this is considered a more
advanced kind of analysis so you should read the manual as well as
some of the literature to become more familiar with it (there are tons
of examples out there).

-Dan

-
-------------------------
Daniel R. Roe
Laboratory of Computational Biology
National Institutes of Health, NHLBI
5635 Fishers Ln, Rm T900
Rockville MD, 20852
https://www.lobos.nih.gov/lcb

_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Tue Oct 25 2016 - 05:30:05 PDT
Custom Search