Dr. Case, I really appreciate how you explained the problem with decomposition schemes, even if the keyword was implemented.
You saved me from potentially falling into that pitfall as a newcomer!
Best,
Matthew
> On Aug 25, 2022, at 9:04 AM, David A Case <david.case.rutgers.edu> wrote:
>
> *Message sent from a system outside of UConn.*
>
>
> On Wed, Aug 24, 2022, Matthew Guberman-Pfeffer via AMBER wrote:
>>
>> I’m wondering if the idecomp output is available in PMEMD for thermodynamic
>> integration calculations. The manual makes it clear that it’s available in
>> SANDER, but I wanted to check if it’s also available in PMEMD. If not, is
>> there another way to get a per-residue decomposition of dv/dl?
>
> You can see in Section 21.2 of the Reference Manual ("Functionality" in
> pmemd) that idecomp!=0 is not supported in pmemd. And, don't be afraid to
> experiment: you'll get the same message if you try to run pmemd with idecomp
>> 0.
>
> Bear this in mind: Although the free energy change (obtained by integrating
> the average of dv/dl) is independent of path that lambda takes in going from
> 0 to 1, the individual pieces of it are not. So a decomposition is not
> well-defined. It may have some qualitative or heuristic use, however.
>
> ....dac
>
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Thu Aug 25 2022 - 06:30:05 PDT
