Pairwise RMSD of a Macro-state MSM network

[daviddaileo]

Hi,

After using 'BuildMSM.py', a macro-state network is obtained. Let's say 100-node network.I am wondering how to extract the centroid structures. Or is there a built-in function which calculates the pairwise RMSD of all the centroid structures?

Thanks,

[kyleabeauchamp]

Use MDTraj to load Gens.lh5 and calculate pairwise RMSD:

http://mdtraj.org/latest/examples/clustering.html

[kyleabeauchamp]

Oops, I misread your question. You'll also have to find the centroids of each each state. That also could be done, but it's not automated so you'll have to write some python.

[kyleabeauchamp]

Or you could do something like drawing random samples from each macrostate, then, calculating RMSDs between the drawn samples.

[daviddaileo]

I prefer the centroid structures. Could you tell me what function extracts centroid structures?

[kyleabeauchamp]

I don't think we have that function written.

On Tue, May 20, 2014 at 5:14 PM, Wei Dai [email protected] wrote:

I prefer the centroid structures. Could you tell me what function extracts
centroid structures?

—
Reply to this email directly or view it on GitHubhttps://github.com//issues/418#issuecomment-43684095
.

[daviddaileo]

Could you show me some pointer to find centroids? Since I've no idea.

[rmcgibbo]

I made a little example of how to do this: http://mdtraj.org/latest/examples/centroids.html

[daviddaileo]

Hi,

The link you sent was very helpful for finding centroids of a micro-state network. What if a macro-state network or MSM network is built from the clustering using PCCA+?

Thanks,