Re: [AMBER] question about dihedral angles

From: Bill Ross <ross.cgl.ucsf.edu>
Date: Wed, 2 Dec 2020 19:20:38 -0800

Stupidly insulting btw, since Carlos's answer was spot on.

BR

On 12/2/20 7:15 PM, Bill Ross wrote:
> Sounds like you might better stick with the compsci part of your program
> if you're going to be insulting professors and really a whole community
> of researchers in compchem, who will be reviewing any papers you write.
>
> Bill
>
>
> On 12/2/20 4:01 PM, Gordon Richard Chalmers wrote:
>> I know you people, Carlos, from SB NY. You are supposed to work first
>> and then talk. Figure it out before saying it is possible or not.
>>
>> Gordon Chalmers
>> ________________________________
>> From: Gordon Richard Chalmers <gordoncs.uga.edu>
>> Sent: Wednesday, December 2, 2020 6:08 PM
>> To: AMBER Mailing List <amber.ambermd.org>
>> Subject: Re: [AMBER] question about dihedral angles
>>
>>
>> By the way, are you sure the angular rotation is about the
>> 2 middle atoms in the 4 bond. That would be good to know if
>> writing your own code.
>>
>> Gordon Chalmers
>> ________________________________
>> From: Gordon Richard Chalmers <gordoncs.uga.edu>
>> Sent: Wednesday, December 2, 2020 5:58 PM
>> To: AMBER Mailing List <amber.ambermd.org>
>> Subject: Re: [AMBER] question about dihedral angles
>>
>>
>> Why don't you try this then: given phi psi angle information of
>> pdb and a linear peptide of its form, reconstruct the pdb file.
>> The problem is not due to numerical instability of the phi/psi
>> angle computation; phi/psi angles are important physically.
>>
>> Gordon Chalmers
>> ________________________________
>> From: Carlos Simmerling <carlos.simmerling.gmail.com>
>> Sent: Wednesday, December 2, 2020 5:52 PM
>> To: AMBER Mailing List <amber.ambermd.org>
>> Subject: Re: [AMBER] question about dihedral angles
>>
>> [EXTERNAL SENDER - PROCEED CAUTIOUSLY]
>>
>>
>> I think that even if you have a list of phi/psi pairs and their values,
>> it's pretty well known that small angle variations can result in large
>> Cartesian changes far from the bond being rotated. I'm not saying this
>> isn't a good plan, but you may run into issues (as you say) with non-planar
>> groups giving small changes in the intended dihedral and leading to
>> significant changes farther down the chain.
>>
>> You should be able to put together a cpptraj script by listing the atom
>> names/residues in dihedral commands to measure them from your pdb, then
>> building a cpptraj script to rotate.
>> Dan's documentation is quite good for academic code standards and I don't
>> think you'll need him to make the inputs for you.
>>
>> tleap can do similar things with commands such as "impose", and in the end
>> it should matter which part gets rotated if you have a single molecule,
>> since all that will be affected is the overall rotational frame of
>> reference. If that is important though, then Dan's suggestion is worth
>> trying. There may be other codes that were intended to build structure
>> models, Amber doesn't really have extensive tools for that purpose.
>>
>> On Wed, Dec 2, 2020 at 5:41 PM Gordon Richard Chalmers <gordoncs.uga.edu>
>> wrote:
>>
>>> Hi Daniel,
>>>
>>> How about giving me a known PDB protein model and a sequence of
>>> rotations to bring its linear form into it? That would be a benefit,
>>> Either
>>> in tleap or cpptraj.
>>>
>>> I think I know what the issue is with non-parallel planes from
>>> the 2 pairs of 3 atoms in the 4 bond of the peptides.
>>>
>>> Gordon Chalmers
>>> ________________________________
>>> From: Daniel Roe <daniel.r.roe.gmail.com>
>>> Sent: Wednesday, December 2, 2020 11:50 AM
>>> To: AMBER Mailing List <amber.ambermd.org>
>>> Subject: Re: [AMBER] question about dihedral angles
>>>
>>> [EXTERNAL SENDER - PROCEED CAUTIOUSLY]
>>>
>>>
>>> Hi,
>>>
>>> The command for calculating dihedral angles in cpptraj is 'dihedral'.
>>> The command for rotating atoms around a dihedral is 'rotatedihedral'.
>>> Both are documented in the manual. The convention in cpptraj is for
>>> all atoms connected to the greater atom index to move; those connected
>>> to the lower atom index are fixed. Cpptraj does the rotation in
>>> Cartesian space - I think tleap may do it via internal coordinates but
>>> I'm not sure about that.
>>>
>>> -Dan
>>>
>>> On Wed, Dec 2, 2020 at 9:04 AM Gordon Richard Chalmers <gordoncs.uga.edu>
>>> wrote:
>>>> Thank you for the help in this.
>>>>
>>>> In a protein the phi and psi angles require 2 adjacent
>>>> residues. Calculating these angles and rotating is
>>>> fairly easy. It is a matter of what is rotated to agree
>>>> to a PDB model starting from a linear backbone model.
>>>>
>>>> I will be working the angle rotations further in tleap today,
>>>> a little delayed at the moment.
>>>>
>>>> In principle, you should rotate the entire peptide chain
>>>> after or before the ith residue as these are bonded in
>>>>
>>>> %% phi C(i-1)-N-CA-C
>>>> %% psi N-CA-C-N(i+1)
>>>>
>>>> The actual rotation angle about the intermediary N-CA
>>>> or CA-C axis should be the dihedral angle geometrically.
>>>>
>>>> Maybe my own Matlab algorithms aren't correct yet, but
>>>> I have checked this many times. I will try to use tleap or
>>>> cpptraj instead as a check.
>>>>
>>>> Gordon Chalmers
>>>> Department of Computer Science
>>>> Complex Carbohydrate Research Center
>>>> University of Georgia
>>>> ________________________________
>>>> From: David A Case <david.case.rutgers.edu>
>>>> Sent: Wednesday, December 2, 2020 8:51 AM
>>>> To: AMBER Mailing List <amber.ambermd.org>
>>>> Subject: Re: [AMBER] question about dihedral angles
>>>>
>>>> [EXTERNAL SENDER - PROCEED CAUTIOUSLY]
>>>>
>>>>
>>>> On Wed, Dec 02, 2020, Karl Kirschner wrote:
>>>>> impose mymolecule {7} { {C1 C2 C3 C4 120.0} }
>>>> One point is this: the command above doesn't specific which atoms are to
>>>> be *moved* in order to make the dihderal be 120 degrees. Should the
>>>> program move C4 (and atoms bonded to it), or move C1 (and atoms bonded
>>>> to it)? Or do something else? This may not be important in some
>>>> workflows, but can be key in others.
>>>>
>>>> The pyMol example has the same problem. Of course, its
>>>> documentation may be better in explaining what choices are made.
>>>>
>>>> These sorts of manipulations are a natural for NAB (aka "molecular
>>>> awk"), but it has its own learning curve and documentation deficiencies.
>>>>
>>>> ....dac
>>>>
>>>>
>>>> _______________________________________________
>>>> 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
>>> _______________________________________________
>>> 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
-- 
Phobrain.com
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Wed Dec 02 2020 - 19:30:03 PST
Custom Search