Re: Advice for Protein Mutation FEP Calculations

From: Brian Radak (
Date: Tue Dec 01 2020 - 13:45:16 CST

alchBondLambdaEnd/alchBondDecouple are indeed niche options (created by
me). In my experience they are only totally necessary when computing
certain relative free energies directly (e.g. pKas, but not hydration or
binding free energies). If you construct your thermodynamic cycle so
that you are computing a relative free difference then they will
probably offer more hindrance than help.



On 12/1/20 2:30 PM, Jérôme Hénin wrote:
> Dear Matthew,
> ----- On 1 Déc 20, at 17:27, Matthew Guberman-Pfeffer wrote:
>> Dear NAMD community,
>> I want to study how one or more residue mutations impact the stability of a
>> protein-protein complex. I've worked through the tutorial on alchemical
>> transformations, but I have some additional questions, and would appreciate
>> advice on how to set options that have since been implemented in NAMD.
>> 1) Should I minimize/equilibrate the bound and unbound protein that will
>> undergo mutations before or after creating the hybrid topology? Does it
>> matter?
> For efficiency, you can equilibrate the initial state alone, then create the hybrid, then do a very quick minimization if needed to ensure the new tolopogy has a correct geometry.
>> 2) How do you choose the value of alchLambdaIDWS?
> You don't have to do it explicitly if you use the FEP scripts in fep.tcl, it is set automatically to the previous lambda value if applicable.
>> 3) When is it appropriate to use alchWCA/alchRepLambdaEnd or
>> alchBondLambdaEnd/alchBondDecouple. If so, again, how do you choose the
>> appropriate settings?
> These are all niche options. If you're unsure what they are, then you don't need them.
> Jerome

This archive was generated by hypermail 2.1.6 : Thu Dec 31 2020 - 23:17:14 CST