pca and correlation network analysis problems

[pel guz]

Hello, I am using the R version 3.4.3 and analyzing my coarse-grained molecular dynamics trajectories with principle component analysis and correlation network analysis tools. Following “Trajectory Analysis with Bio3D”, “Principal Component Analysis” and “Protein Structure Networks with Bio3D” tutorails.
My coarse-grained molecular dynamics trajectories (5000 frames) in .dcd format . I have a few problems with the results as below;

1- with pca.xyz() command I am getting the pcs without any error and when I plotted proportion of variance, I got logical results, in terms of ~70 percent of total variance captured in first 3-4 components. But when I wanted to see the contribution of each residue to the for example first two components, everything looks like the same. Briefly, I can’t see the difference between principle components. The same thing continiues when I wanted to see the pdb format of trajectory by using mktrj.pca(). What could be the reason? Where should I start to check if there is problem with my principle components results?

2- Does it creates a problem if my molecular dynamics trajectory is already in a coarse-grained scale (one bead-Cα) before starting correlation network analysis? When I checked the paper of your group (Scarabelli & Grant, 2014, Biophysical Journal), I understood that residue-wise linear mutual information were calculated from atomic fluctuations as first step to correlation network analysis. So I am confused if it is a problem giving a coarse-grained trajectory as an input because my cna() calculation is continuing for 2 days. Thank you for your help and comments in advance.

[Barry Grant]

There is no problem with using coarse grained trajectories. Just be sure to read a corresponding PDB file for your trajectory with the exact same number of atoms/particles.

The mktrj() PDB trajectory will include the same number of atoms so it is unclear what your issue might be here.

Beyond this it is not clear how we can help you from the amount of information you provide (see the posting guide). In particular, we can not reproduce your possible issue from the text you write.

[Barry Grant]

insufficient information provided