Hi,

I find that the function sample_states() in io.sampling is useful for sampling conformations from each of N microstates in my MSM. Using the following code:

import mdtraj as md

from msmbuilder.io import load_trajs, save_generic, preload_top, backup, load_generic
from msmbuilder.io.sampling import sample_states

## Load
meta, ttrajs = load_trajs('ttrajs')
msm = load_generic('msm.pickl')
kmeans = load_generic('kmeans.pickl')

inds = sample_states(ttrajs, kmeans.cluster_centers_)
top = preload_top(meta)
traj = md.join(
md.load_frame(meta.loc[traj_i]['traj_fn'], index=frame_i, top=top)
for traj_i, frame_i in inds)

1. Am I right in assuming that this picks a set of N "random" conformations each from every microstate in the same order as the rows/columns in the transition matrix?

2. Is there a way for me to verify that each of these conformations adequately represents their respective microstate, such as a metric for RMSD variance in each cluster?

Just FYI, the msm itself was constructed using virtually identical protocol as in http://msmbuilder.org/3.8.0/tutorial.html.

Thanks in advance!