Hi David,
For mutating amino acids for TI softcore, do you know what set of atoms are optimal common atoms? I currently only specify C, N, O, CA, H, HA as common atoms. Some residues do share CB, CG etc. common atoms, but I do not know if there are amino acid specific side chain C-C angles that will limit the specification of these atoms as common. Sometimes the wild type and mutated residue side chains could have very different conformations to avoid clashes, so these non backbone atoms usually have different positions. By specifying non backbone atoms as common atoms, will the later MD conformation sampling of these atoms be affected or if the common atoms can sample different conformations for the wild type and mutated residues?
My test runs showed that the ddG predictions from TI performs poorly when mutating Gly or small side chain residues to large side chain residues with clashes expected at protein interfaces, I wonder if there is any way to improve the protocol. Running for longer simulations seem did not help so far (2ns vs 5ns per window, 12 windows in total).
In general, is it possible to tell the reliability of the calculated ddG from the trajectory or energy outputs of TI MD? I did not see any tutorials on this.
Best,
Zizhang
> On Mar 26, 2021, at 4:56 PM, zz sheng <shengzizhang.gmail.com> wrote:
>
> Thanks David.
> Will try your suggestions. The changes is about 0.1A.
>
> Best,
>
> Zizhang
> From: David A Case <david.case.rutgers.edu>
> Sent: Friday, March 26, 2021 4:40:19 PM
> To: AMBER Mailing List <amber.ambermd.org>
> Subject: Re: [AMBER] Distance tolerance for identifying common atoms in thermodynamics integration
>
> On Fri, Mar 26, 2021, zz sheng wrote:
> >
> >I am preparing the softcore mask for TI simulation. Two copies of the
> >receptor was included in one file with one amino acid mutation for topology
> >file preparation by tleap. It turns out that tleap alters slightly the
> >coordinates of the backbone atoms of the mutated residues which should
> >assumably be identical between the two receptor copies. So now Amber can
> >not identify common backbone atoms because of the coordinate changes. I
> >wonder if there are parameters can be set to adjust the distance tolerance
> >for identification of common atoms in Amber. Or, to disable tleap to change
> >the coordinates?
>
> This is odd: it would be nice if you could create a small example showing
> the problem. What do you mean by the term "slightly"?
>
> Also, try adding the following command at the beginning of your
> tleap script, to see if it helps:
>
> set default nocenter on
>
> But this is a guess: if the WT and mutated atoms have identical coordinates
> in the input structure, that relationship should not be changed by tleap.
>
> ....dac
>
>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber <http://lists.ambermd.org/mailman/listinfo/amber>
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Mon Apr 05 2021 - 08:30:02 PDT
