From: Jérôme Hénin (jerome.henin_at_ibpc.fr)
Date: Wed Dec 02 2020 - 17:11:54 CST
> Jerome— Thanks for the practical advice. I’m starting from a low resolution
> CryoEM structure, so I’ll take your suggestion of relaxing, then building the
> hybrid. Regarding fep.tcl, how will the script work if the job times out on an
> HPC and I need to restart it?
The script expects all windows to have the same number of timesteps, so the simplest would be to run an integer number of windows in each job, or discard any incomplete window at the end.
I you use IDWS you need to specify the previous lambda value when restarting (see comments in fep.tcl for an example).
> I have a further question I’d greatly appreciate all of your thoughts on. Apart
> from my examination of mutants, I need to compare the stability of an A-A or an
> AB-AB complex (i.e. the association free energy for two A proteins, or two AB
> protein dimers). Each protein unit (A or B) is about 150 residues. I care about
> the relative (not absolute) free energies only. How can this problem be tackled
> with the free energy methods in NAMD?
On this scale, I think you're better off looking at association/dissociation through space. I don't see a good method to look at relative free energies for changes of this type. It's a very challenging problem.
This archive was generated by hypermail 2.1.6 : Fri Dec 31 2021 - 23:17:10 CST