nileshjchoudhary / sgoop Goto Github PK

Any publications using this code are requested to cite the following references:
1. Tiwary and Berne, Proc. Natl. Acad. Sci. 113 DOI: 10.1073/pnas.1600917113
2. Smith, Pramanik, Tsai and Tiwary, J. Chem. Phys., 2018, 149 234105; DOI: 10.1063/1.5064856
3. Tiwary and Parrinello, J. Phys. Chem. B, 2015, 119 736; DOI: 10.1021/jp504920s

Use reweight to convert biased trajectories to usable unbiased probability distributions and analyze them with SGOOP.
SGOOP may also be used with unbiased data using the rc_eval function.

Run SGOOP on multiple numbers of wells (sgoop.wells) checking each time that the number is self-consistent with the number
of wells in the probability distribution (depth > kT).  The optimized RC with highest number of self-consistent wells is
the optimal RC.  For instance, if 2 wells was self-consistent but 3 was not the 2 well RC would be the optimum.  These
consistency checks will soon be automated and sgoop.wells values will be looped through.

Files needed: Maximum Caliber (unbiased) trajectory, COLVAR file, FES files


The following function is an example of using SGOOP and reweight together.  This would be used with an optimization method
such as simulated annealing:

import sgoop
import reweight as rw
import scipy.optimize as opt
import numpy as np

def opti_func(rc):
    p = rw.reweight(rc) # Calculating probability along RC
    sgoop.set_bins(rc,np.shape(p)[0],rw.s_min,rw.s_max) # Setting bins for maxcal
    return sgoop.sgoop(rc,p)   # Calculating the spectral gap on the given RC.