Allow override t0 and g_t when computing free energy #585
Conversation
|
@ijpulidos @mikemhenry : What do you think of my changes here? Is there a better way to allow the user to override these parameters? |
|
Few ideas:
# Given a reporter, create an analyzer
analyzer = MultiStateSamplerAnalyzer(reporter, max_n_iterations=max_n_iterations)
# Compute free energy
f_ij, df_ij = analyzer.get_free_energy()Another option would be to add some setters to analyzer so you could do something like: # Given a reporter, create an analyzer
analyzer = MultiStateSamplerAnalyzer(reporter, max_n_iterations=max_n_iterations)
# Override t0 and g_t
t0 = 1000
subsample_rate = 10
analyzer.set_subsample_rate(subsample_rate)
analyzer.set_t0(t0)
# Compute free energy
f_ij, df_ij = analyzer.get_free_energy()Really it all depends on if |
|
Re 1: I've already looked into ways to improve the way its automatically determined by replacing the potential energy timeseries that's fed in to Re 2: Yes, the normal calls are the ones you specified in the first code snippet. I just added the call to |
|
Make If the user specifies all of them, don't call |
Description
The
MultiStateAnalyzerautomatically determines t0 and g_t here.I have discovered that for my barnase:barstar repex simulations, t0 is always 1, which is too small. I'd like to be able to manually feed in t0 (and g_t) before retrieving the free energy.
This PR allows overriding t0 and g_t in the following way:
Todos
Status