R.prep_observable(lam=lam, free_energy=energies[i],
filename=File, dloggamma=np.log(1.01),gamma_min=0.2,gamma_max=5.0)
elif File.endswith('pf'):
R = Restraint_pf('8690.pdb',ref='gaussian')
R.prep_observable(lam=lam, free_energy=energies[i],
filename=File)
ensemble[-1].append(R)
print ensemble
##########################################
# Next, let's do some posterior sampling
########## Posterior Sampling ############
sampler = PosteriorSampler(ensemble)
sampler.compile_nuisance_parameters()
sampler.sample(nsteps) # number of steps
print 'Processing trajectory...',
sampler.traj.process() # compute averages, etc.
print '...Done.'
print 'Writing results...',
sampler.traj.write_results(os.path.join(outdir,'traj_lambda%2.2f.npz'%lam))
print '...Done.'
sampler.traj.read_results(os.path.join(outdir,'traj_lambda%2.2f.npz'%lam))
print 'Pickling the sampler object ...',
for i in range(energies.shape[0]):
print
print '#### STRUCTURE %d ####'%i
if verbose:
print data[i]
s = Restraint('8690.pdb',lam,energies[i],data = data[i])
ensemble.append( s )
##########################################
# Next, let's do some posterior sampling
########## Posterior Sampling ############
sampler = PosteriorSampler(ensemble)
sampler.sample(nsteps) # number of steps
print 'Processing trajectory...',
sampler.traj.process() # compute averages, etc.
print '...Done.'
print 'Writing results...',
sampler.traj.write_results(os.path.join(outdir,'traj_lambda%2.2f.yaml'%lam))
print '...Done.'
# pickle the sampler object
print 'Pickling the sampler object ...',
outfilename = 'sampler_lambda%2.2f.pkl'%lam
print outfilename,
fout = open(os.path.join(outdir, outfilename), 'wb')
plt.plot(exp_distances, model_distances,'.')
plt.plot([0,5],[0,5],'k-')
plt.xlabel('exp distance ($\\AA$)')
plt.ylabel('model distance ($\\AA$)')
plt.title('model %d'%lookat[i])
plt.show()
##########################################
# Next, let's do some posterior sampling
else:
#sampler = PosteriorSampler(ensemble, use_reference_prior=True, sample_ambiguous_distances=False)
sampler = PosteriorSampler(ensemble, dlogsigma_noe=np.log(1.02), sigma_noe_min=0.05, sigma_noe_max=5.0,
dlogsigma_J=np.log(1.02), sigma_J_min=0.05, sigma_J_max=20.0,
dloggamma=np.log(1.01), gamma_min=0.5, gamma_max=2.0,
use_reference_prior=not(args.noref), sample_ambiguous_distances=False)
#sampler = PosteriorSampler(ensemble, use_reference_prior=True)
sampler.sample(args.nsteps) # number of steps
print 'Processing trajectory...',
sampler.traj.process() # compute averages, etc.
print '...Done.'
print 'Writing results...',
sampler.traj.write_results(os.path.join(args.outdir,'traj_lambda%2.2f.yaml'%args.lam))
print '...Done.'
# pickle the sampler object
print 'Pickling the sampler object ...',
s.build_groups()
# Compare the model and exp distances
for s in ensemble:
print s
for d in s.distance_restraints:
print 'd.model_distance =', d.model_distance,
print 'd.noe_distance = ', d.noe_distance
print 's.distance_equivalency_groups', s.distance_equivalency_groups
print 's.distance_ambiguity_groups', s.distance_ambiguity_groups
print '-------'
##########################################
# Next, let's do some posterior sampling
sampler = PosteriorSampler(ensemble)
# set sampling parameters
sampler.sigma_noe = 0.1
sampler.dsigma_noe = 0.005
#sampler.sigma_Jcoupl = 1.0
#sampler.dsigma_Jcoupl = 0.2 # controls the magnitude of steps in sigma_Jcoupl
sampler.sample(100000) # number of steps
sampler.traj.plot_results()
"""
ATOM 1 C1 UNK 1 1.484 2.515 0.317 1.00 0.00 C
ATOM 2 O1 UNK 1 1.764 3.456 -0.687 1.00 0.00 O
