[gmx-users] Re: help with chromophore of a GFP

Mark Abraham mark.j.abraham at gmail.com
Thu Apr 4 22:05:17 CEST 2013


On Thu, Apr 4, 2013 at 11:40 AM, Anna Marabotti <amarabotti at unisa.it> wrote:

> Dear gmx-users, dear Mark,
> thank you for all the help you are giving me for this subject. I'm still
> proceeding to my target..
> Thanks to your previous suggestions, I was able to parametrize bonds and
> angles of CFY. Now I'm dealing with dihedrals...
>
> In the parameter files derived by Antechamber I have 2 different sections
> for dihedrals and impropers. They are characterized by these values:
>
> DIHEDRALS:
> The actual torsion potential is (PK/IDIVF)* (1 + cos(PN*phi - PHASE). The
> components should be those reported in the column of the file, i.e.:
>
> - first column: IDIVF: a factor by which the torsional barrier is divided
> - second column: PK: the barrier height divided by a factor of 2
> - third column: PHASE: the phase shift angle in torsional function (in deg)
> - fourth column: PN: the periodicity of the torsional barrier
>
> An example for this section is:
>
> DIHE
>
> NT-CT-C -O10.000180.0002.000same as X -c -c3-X
>
> NT-CT-C -N10.000180.0002.000same as X -c -c3-X
>
> NT-CT-CT-CT10.1560.0003.**000same as X -c3-c3-X
>
> NT-CT-CT-HC10.1560.0003.**000same as X -c3-c3-X
>
> CT-C -N -CT12.500180.0002.000same as X -c -n -X
> ..............................**..
>
>
> IMPROPERS:
> The input is the same as for dihedrals except that the torsional barrier
> height is NOT divided by IDIVF. Therefore the components are:
>
> - first column: PK: the barrier height divided by a factor of 2
> - second column: PHASE: the phase shift angle in torsional function (deg)
> - third column: PN: the periodicity of the torsional barrier
> An example is:
>
> IMPROPER
>
> CT-N -C -O1.1180.02.0Using default value
>
> C -CT-N -H1.1180.02.0Using default value
>
> CK-CT-CM-NB1.1180.02.0Using default value
>
> C -CC-CC-NB1.1180.02.0Using default value
>
> ..........................
>
> I should report these information in ffbonded.itp in order to fix the
> values for dihedrals for CFY. However, I have some difficulties in doing
> it, and description in Table 5.5 of the manual is rather cryptic to me.
> First of all, I don't understand why in ffbonded.itp there are two
> separated [dihedral types] sections, I see that in the first section the
> column that I presume corresponds to "f.tp" in Table 5.5 (but in
> ffbonded.itp it is called "func"), is all equal to 4, whereas in the second
> section this value is 9 for all.
>
> An example for the fist section is:
> [ dihedraltypes ]
> ;i  j   k  l     func      phase      kd      pn
> CA  CA  CA  OH       4      180.00     4.60240     2     ; new99
> H5  O   C   OH       4      180.00     4.60240     2     ; new99
> H5  O   C   OS       4      180.00     4.60240     2     ; new99
> CM  CT  CM  HA       4      180.00     4.60240     2     ; new99
> CA  CA  CA  Br       4      180.00     4.60240     2     ; new99
> ..............................**...
>
> and for the second section:
>
> [ dihedraltypes ]
> ;i   j   k   l       func
>  CT  CT  OS  CT    9       0.0      1.60247     3  ;
>  CT  CT  OS  CT    9     180.0      0.41840     2  ;
>  C   N   CT  C     9       0.0      1.12968     2  ; new for 99sb
>  C   N   CT  C     9       0.0      1.75728     3  ; new for 99sb
>  N   CT  C   N     9     180.0      1.88280     1  ; new for 99sb
>
> If I understand correctly, then this means that the first section is for
> periodic improper dihedrals, and the second section is for proper dihedral
> multi, right?


Yes.


> I don't understand the difference between "improper dih." f.tp. 2 and
> "periodic improper dih." f.tp.4.


The parameters are for a different mathematical function, described in
chapter 4. If you check out the 4.6 manual, I've introduced
cross-references to those sections. The parameter names in Table 5.5 are
common with the equations, and the ordering in the .itp files matches that
implied by Table 5.5. Unfortunately the .itp files are an inconsistent
mess, and can't use the same subscript pattern as the equations anyway. But
AFAIK you should regard the .itp comments as indicative and not definitive.

Moreover, if I can imagine that PHASE and PN mean the same in both cases
> (ffbonded.itp and Antechamber values), I don't understand if kd is
> corresponding to PK, or to the result of (PK/IDIVF)* (1 + cos(PN*phi -
> PHASE).
>

You need to lay out the relevant equations and transform the parameters
that work with the AMBER equations to work in the functions that the
function types implemented by GROMACS. If you want to be sure of how
something is working, you should compute some
http://www.gromacs.org/Documentation/How-tos/Single-Point_Energy values.
Set the k<whatever> value for a bondedtype under test to zero to establish
the baseline. Then set it to the value you think is right and inspect the
difference with g_energy on the .edr file. Then double k and verify the
energy difference agrees with what you expect. Unfortunately, there is no
automated version of this process. Building a small peptide system as a
test bed can be an option.

Using a tool like https://code.google.com/p/acpype/ might make the job
easier. Unfortunately, nobody has yet taken the time to build something
that can handle all Antechamber-to-GROMACS conversion tasks that a user
could want.

Mark

Could you please give me some hints on how I can edit correctly the
> ffbonded.itp file using data obtained by Antechamber?
>
> Thanks again and best regards
> Anna
> --
> gmx-users mailing list    gmx-users at gromacs.org
> http://lists.gromacs.org/**mailman/listinfo/gmx-users<http://lists.gromacs.org/mailman/listinfo/gmx-users>
> * Please search the archive at http://www.gromacs.org/**
> Support/Mailing_Lists/Search<http://www.gromacs.org/Support/Mailing_Lists/Search>before posting!
> * Please don't post (un)subscribe requests to the list. Use the www
> interface or send it to gmx-users-request at gromacs.org.
> * Can't post? Read http://www.gromacs.org/**Support/Mailing_Lists<http://www.gromacs.org/Support/Mailing_Lists>
>



More information about the gromacs.org_gmx-users mailing list