On Thu, Nov 16, 2017 at 5:27 AM, Lizelle Lubbe <LBBLIZ002.myuct.ac.za> wrote:
>
> Firstly, I would like to ask if it would be meaningful to calculate the principal motions for the :protein&CA and the :glycans&!.H= separately. The protein is known to hinge
> The PCA analysis I performed yesterday was of the :glycoprotein&!.H= and in NMWiz it predominantly shows glycan motions but the porcupines for protein motion makes the output cluttered.
It's fine to do, it will just tell you something different. It all
depends on what your reference frame is. For example, if you RMS-fit
on your protein but then perform Cartesian PCA on the glycans, the
principal components will reflect the motion of the glycans with
respect to the protein, etc.
> Secondly, which principal component does the .nmd file contain? If it is showing all 3 how can I separate them into a .nmd file for each individual component?
It contains all the eigenvectors, or up to nmwizvecs. I'm not sure
that you can pick out specific ones but I think that the NMwiz plugin
from VMD allows you to do that, right?
>
> Lastly, I would like to display the covariance matrix as a heatmap and the PCA results as scree plots. I previously used Bio3D in R but don't have enough RAM for the current MD analysis.
> Could these plots be generated in cpptraj from the output of the above script?
Sure - just use the 'projection' command to project your coordinates
along your principal components and then the 'hist' command on the
datasets generated by 'projection'. I think there are examples in the
manual and also here:
http://www.amber.utah.edu/AMBER-workshop/London-2015/pca/
Hope this helps,
-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 Thu Feb 01 2018 - 06:30:04 PST
