From: Josh Vermaas (
Date: Tue Mar 04 2014 - 20:28:34 CST

Hi Joe,

Ahh... What you are asking for is unfortunately beyond what is easily
accessible/currently available that I know of. Making a in silico
mutation (Proline to His) is easy (just replace the residue wholesale).
Calculating the free energy change of such a mutation is MUCH harder,
and I'm not sure if its a well-defined quantity. Proline, as I'm sure
you are aware, is unique, in that it links back on itself. This in
itself isn't a big problem, except for how it influenced the way it was
parameterized. For starters, the charge on the backbone nitrogen and
alpha carbon are different and they have a different atomtype, which
means that the alchemy no longer is restricted to the sidechain (an
assumption the mutator plugin exploits to simplify the code and the
topology of the hybrid residues).

What is probably simplest is to make a psf/pdb pair where BOTH Pro and
His are present and connected to the adjacent residues. This basically
breaks your protein into two pieces, joined together by an alchemical
section. This would be technically correct, however since the phi/psi
for prolines are unique, I don't think you'll get convergence on any
kind of acceptable timescale, since the two different connectors will
try to enforce two different conformations on the rest of the protein.
In short, there is a reason this isn't explicitly supported by mutator,
since the results you get are not meaningful per se.

Good luck!
-Josh Vermaas

On 03/04/2014 07:53 PM, Daou, Joseph A wrote:
> Thank you for the help, Josh!
> Here is another question.
> I tried to turn the Free energy perturbation calculation with the
> mutation of Pro--> X (X=any amino acid). In my case X is H
> (Histidine). NAMD2 could not recorgnize P2H topology. Do you have a
> hybrid topology for Pro--> H mutation to run Free energy perturbation?
> If not, can I generate such a topology file by myself?
> Thanks,
> Joe
> ------------------------------------------------------------------------
> *From:* Josh Vermaas <>
> *Sent:* Tuesday, March 04, 2014 4:39 PM
> *To:* Daou, Joseph A;
> *Subject:* Re: vmd-l: Mutate Residue Extension
> Hi Joseph,
> Have you tried running mutator from the tkconsole? If so, what is the
> error message exactly? Mutating from a proline to anything else is
> procedurally no different than mutating any other residue. What might
> cause a problem in VMD 1.9.1 is if the resid for the residue has an
> insertion code in it (48A instead of just 48), which would make psfgen
> in that version very very unhappy. With the exact error message, it
> would be easier to pin down exactly what is going on.
> FYI, the mutate residue plugin is just a frontend for psfgen, so what
> it does is basically the following:
> #psfgen setup
> package require psfgen
> topology charmblahblahblah.rtf
> #Now build the psf/pdb pair (I'm assuming your protein has one
> segment, with the mutation occuring at resid 123)
> resetpsf
> segment PRO {
> pdb protein.pdb
> mutate 123 whateveryouwantthatisntproline
> }
> coordpdb protein.pdb
> guesscoord
> regenerate angles dihedrals
> writepsf new.psf
> writepdb new.pdb
> So if it comes down to it, you could play around with a script like
> that to build it however you want.
> -Josh Vermaas
> On 03/04/2014 02:25 PM, Daou, Joseph A wrote:
>> Dear VMD specialists,
>> I am an undergraduate Biochemistry researcher at the University of
>> New Haven. I had a question regarding the "Mutate Residue" extension.
>> I am having trouble mutating the Proline residues on my protein to
>> another amino acid. The mutator cannot recognize where proline is on
>> a specific protein we are studying. Do you have any suggestions on
>> other methods to try? I would even greatly appreciate any advice on
>> this issue.
>> Thank you in advance for the help!
>> Best,
>> Joseph