[gmx-users] D- &/or L- cmap for charmm36

Justin Lemkul jalemkul at vt.edu
Thu Oct 13 04:07:11 CEST 2016



On 10/12/16 9:59 PM, HENRY WITTLER wrote:
> Thanks for previous replies Justin.
>
> Since Im at the end of my PhD and the D-aa simulation is just an extra to the PhD-project. I can keep the charmm36 D-aa files etc, if I need to continue this later,
> so I have the old cmap parameters to match my old gmx. 5.0.4. charmm36-mars2014 L-aa simulation.

Please note, as the comment in the CHARMM force field file notes, CMAP has 
changed as a result of the latest 2016 paper (in press).  I will do a full 
conversion and announce the port here when the paper is available online as part 
of a new version of CHARMM36 that will include some other things.

I would not recommend trying to use D-CMAP terms from the file below with the 
March 2014 (very outdated!) version of the CHARMM36 port.

> Not sure if I can dig deeper to answer my next question, if it is something I can solve myself I dont expect a reply.
> The CMAP energy unit conversion from CHARMM to GMX, do not seem straightforward to me, since the format is different, see below.  Is there a script to do this conversion readily?

It's rather simple.  It's a 24 x 24 matrix.  GROMACS specifies this with "24 24" 
and the first 24 terms are the entries for phi = -180, the next 24 for phi = 
-165, etc.

-Justin

>
> GMX
> ; This force field generated by charmm2gmx.py from
> ; multiple charmm parameter files
> ; and multiple charmm topology files
> [ cmaptypes ]
> C NH1 CT1 C NH1 1 24 24\
> 0.53048936 3.21624080 4.06375184 5.23405848 8.87430584 11.38227912 8.74221696 7.48848136 3.26716008 -2.88057522\
> 4.18872792 -9.20697568 -20.19897128 -20.17293425 -20.55692503 -15.02518750 -11.57481006 -11.64909280 -10.27000038 -9.81981034\
> -9.77110440 -6.37079270 -3.98125173 -0.15334360 -0.53192447 5.76174456 6.59825168 7.83366136 10.03737416 10.40405992\
> 10.19537874 8.07122888 4.54573190 2.69198560 1.08230038 -11.71704096 -16.77565177 -17.30212030 -14.30965656 -10.88735376\
> -9.61955155 -6.28119234 -4.60984752 -3.60423986 -2.67846291 -0.86902098 -4.50342330 -4.69457352 0.35174470 5.94260633\
>
> CHARMM
> CMAP
> ! 2D grid correction data.
> ! Finalfix3, Feig/Best/MacKerell 2010
>
> ! Jing Huang/Alex MacKerell adjustments to correct for
> ! oversampling of alpha L conformation.  2016/1
>
> !Direct conversion using Jared Ostmeyer/Benoit Roux flip_cmap.py script
> !to invert L CMAP to create D CMAP.  Proline unchanged.
>
> ! D-alanine map--
> ! note the new type for the alpha carbon
> C    NH1  CTD1  C    NH1  CTD1  C    NH1   24
>
> ! phi = -180.0
>          0.13         -0.04         -0.95         -1.52         -2.34
>         -2.35         -2.45         -2.78         -2.77         -3.59
>         -4.91         -4.82         -4.83         -2.20          1.00
>         -0.69          0.78          1.79          2.09          2.72
>          2.12          1.25          0.97          0.77
>
> ! phi = -165.0
>         -0.81         -1.81         -2.75         -3.25         -3.55
>         -3.56         -3.84         -3.67         -4.19         -5.14
>         -5.89         -6.41         -4.56         -0.74         -2.74
>         -1.09          0.63          1.52          1.70          1.64
>          1.28          0.54          0.38         -0.07
>
> Cheers,
> Henry
>
> On 10/4/16 2:00 AM, HENRY WITTLER wrote:
>>
>> Thanks, for the previous reply.
>>
>> Will the CHARMM cmap and topology for D-amino acids be included in gmx for a later version of gmx charmm36?
>
> It doesn't seem straightforward to convert the charmm D-aa cmap parameters to gmx cmap.itp format.
> Can this program be used for this 'cgenff_charmm2gmx.py' at http://mackerell.umaryland.edu/charmm_ff.shtml#charmm
>>
>
> I will consider putting it in.  Note that we have an update to the CHARMM36
> protein parameters (specifically CMAP) that is about to be published, so no
> update will be posted until then (though the files, in CHARMM format, are
> available, but we're waiting on the conversion until the reference is
> available).  The D-amino acid CMAP, however, has not been adjusted, but can
> easily be converted.
>
> In the meantime, you can certainly proceed with the existing files.  The CMAP
> conversion is just converting energy units.  And a new DSER entry would just
> require changing the CA type in the existing SER .rtp entry to create a new residue.
>
> -Justin
>
>>
>> Cheers,
>> Henry Wittler
>>
>> Phd in Molecular modelling group (Brian J. Smith)
>> Department of Chemistry and Physics, La Trobe Institute for Molecular Science, La Trobe University, Victoria 3086, Australia
>> Tel: 0432901627
>>
>>
>>
>> Message: 3
>> Date: Wed, 21 Sep 2016 15:08:23 -0400
>> From: Justin Lemkul <jalemkul at vt.edu<https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users>>
>> To: gmx-users at gromacs.org<https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users>
>> Subject: Re: [gmx-users] D- &/or L- cmap for charmm36
>> Message-ID: <5156078a-2572-401b-ec29-62dd5ae8bb6f at vt.edu<https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users>>
>> Content-Type: text/plain; charset=windows-1252; format=flowed
>>
>>
>>
>> On 9/21/16 1:43 AM, HENRY WITTLER wrote:
>>> Greetings.
>>>
>>>
>>> If anyone can give insight please.
>>>
>>>
>>> I want to do a MD for insulin a 51aa protein, mutating one residue from L to D-Ser.
>>>
>>> I have runned 1.5us for the L- Ser (with cmap), and want to repeat it for the D-Ser, with gmx v.5.0.4 charmm36 mars14 version.
>>>
>>>
>>> It seems in the traditional CHARMM one have to change dihedrals involved with changing L- to D-
> (http://www.ks.uiuc.edu/Training/Tutorials/science/topology/topology-html/node4.html), also the cmap is for L-aa
> (http://mackerell.umaryland.edu/~kenno/cgenff/faq.php<http://mackerell.umaryland.edu/%7Ekenno/cgenff/faq.php<http://mackerell.umaryland.edu/%7Ekenno/cgenff/faq.php%3Chttp://mackerell.umaryland.edu/%7Ekenno/cgenff/faq.php>>).
>>>
>>>
>>> In link http://www.swisssidechain.ch/D_residues.php gives topology D-aa for gmx-charmm, without changing anything about cmap.itp,
> only .hdb and .rtp.
>>>
>>>
>>> Previous discussions (http://gromacs.org_gmx-users.maillist.sys.kth.narkive.com/r1t0ZqeF/converting-l-to-d-amino-acid-in-the-charmm-force-field-in-gromacs-where-to-alter-dihedral)
> and my own testing MD and looking at it in vmd etc, seems to imply that the .top, .itp do not distinguish about chirality, is this the case?
>>>
>>> How is cmap.itp implemented in gmx, is these parameters for L-, and D- aminoacids?
>>>
>>
>> The parameters provided are for L-amino acids.  There are D-amino acid CMAP
>> parameters available in CHARMM, but not in GROMACS format.  See the latest force
>> field distribution on http://mackerell.umaryland.edu/charmm_ff.shtml#charmm and
>> look in stream/prot/toppar_all36_prot_d_aminoacids.str
>>
>> The D vs L stereochemistry is made possible by assigning a special atom type to
>> the CA atom (CTD1 instead of CT1) so you can apply residue-specific parameters.
>>
>> -Justin
>

-- 
==================================================

Justin A. Lemkul, Ph.D.
Ruth L. Kirschstein NRSA Postdoctoral Fellow

Department of Pharmaceutical Sciences
School of Pharmacy
Health Sciences Facility II, Room 629
University of Maryland, Baltimore
20 Penn St.
Baltimore, MD 21201

jalemkul at outerbanks.umaryland.edu | (410) 706-7441
http://mackerell.umaryland.edu/~jalemkul

==================================================


More information about the gromacs.org_gmx-users mailing list