Computing residency times for pcoords/auxdata within range

[Gabriel Monteiro da Silva]

Hi,

I was wondering if there is an easy way to calculate residency times for walkers in a given range of observables for a WESTPA simulation (i.e., for how many iterations each walker stays in that range). 

Here's a figure for clarity. I'm trying to measure if an intermediate conformation (defined by dihedral angles, hence the wrapping) is short-lived compared to others in an equilibrium simulation, but that's not immediately clear from the probabilities plot.

Many thanks!

[Daniel Zuckerman]

Gabriel, you could try building a Markov model (really haMSM if your sim started from a localized region) and use that to calculate relaxation times of auto-correlation functions.  So, for instance, if you can define state B as a set of microbins in the (ha)MSM, you could calculate

< h_B(0) h_B(t) > 

where h_B is an indicator function.  The associated relaxation time would be a reasonable proxy for the lifetime.

--Dan

---

From: westpa...@googlegroups.com <westpa...@googlegroups.com> on behalf of Gabriel Monteiro da Silva <gabriel.m...@gmail.com>
Sent: Thursday, November 11, 2021 8:35 PM
To: westpa-users
Subject: [EXTERNAL] [westpa-users] Computing residency times for pcoords/auxdata within range

 

--
You received this message because you are subscribed to the Google Groups "westpa-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to westpa-users...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/westpa-users/03d60bbb-0370-4306-8435-b8725941f3fdn%40googlegroups.com.

[Gabriel Monteiro da Silva]

Thanks Dan, I'll try that!