Dear Sir,
>yes, we do this very often. the important thing is to be aware of the effect
>>>of fitting. for example, it's best to fit the entire protein and then do
RMSD for each residue without refitting.
I got confused, please explain (and please correct me where I am wrong)
I was NOT saying about rmsd fitting (which tries to superimpose two pdbs to have a least value of rmsd)
but I meant rmsd only (distance between equivalent points).. e.g. C1 of 1ns.pdb vs C1 of 2ns.pdb.. Does AMBER performs rmsd fitting?my previous mail was pointing towards rmsd NOT fitting..
>for example, it's best to fit the entire protein and then do
>RMSD for each residue without refitting
this arises another question in me, the fitting method may cause NON Equivalent points to superimpose to give least rmsd value(whole protein rmsd fitting).... which will change 3D coordinates of one pdb compared to other (yes, intra-structure points w.r.t each other donot change). So now if we calc. per residue basis rmsd will not be good to say with CONFIDENCE. Also, Even if coordinates are not changed (if my perception is wrong) then why we should give chance for NON equivalent points to get superimposed to give least rmsd fitting value. I mean to say we should consider equivalent points only.
if only rmsd of equivalent points is calculated then only with CONFIDENCE we can say about superposition or nonsuperposition of similar points (in this we should directly do rmsd, not fitting, on residue basis).
regards,
JIomm
--- On Sat, 3/20/10, Carlos Simmerling <carlos.simmerling.gmail.com> wrote:
From: Carlos Simmerling <carlos.simmerling.gmail.com>
Subject: Re: [AMBER] general query:RMSD
To: "AMBER Mailing List" <amber.ambermd.org>
Date: Saturday, March 20, 2010, 4:41 PM
yes, we do this very often. the important thing is to be aware of the effect
of fitting. for example, it's best to fit the entire protein and then do
RMSD for each residue without refitting. if you fit each residue before
calculating the rmsd, the result is less meaningful.
On Fri, Mar 19, 2010 at 4:21 PM, Jio M <jiomm.yahoo.com> wrote:
> Dear AMBER users,
>
> I have a general query regarding RMSD... why we not carry out rmsd on
> residue basis e.g. on per amino acid basis. (sorry if same discussion had
> been occurred earlier)
>
> I mean to say, RMSD is ''sort of averaging'' out the things as it is being
> divided by n number of atoms. So if (less or more) number of atoms are much
> deviated then it might not be clear (that is we can miss that deviation of
> atoms) from whole protein rmsd comparison.
>
> It can easily be determined on basis on per amino acid basis rmsd
> comparison or we can say on small number of atoms say 10 (obviously more
> less number of atoms being considered more exact we will reach). It will
> help to know which group is exactly responsible for variation even if it was
> not seen for WHOLE protein rmsd calculation (being averaged out)
>
> hope it is clear.
>
>
> thanks and regards,
>
> JIomm
>
>
>
>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
>
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Sun Mar 21 2010 - 02:30:03 PDT
