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

Mark Abraham mark.j.abraham at gmail.com
Wed Apr 10 20:40:04 CEST 2013

On Wed, Apr 10, 2013 at 5:25 PM, Anna Marabotti <amarabotti at unisa.it> wrote:

> Dear Mark,
> I am really exhausted about this story. I read your message below, and if
> I understand correctly, nothing is apparently wrong in what I'm doing,
> however things are not going at all.
> Following your suggestions, I removed in ffbonded.itp the lines defining
> the CT-CT-N-C dihedral, as well as the other duplicated dihedrals that I
> added. I checked carefully and didn't find any other duplication that could
> be referred to the error in the dihedral that was already set.
> In your message below, you told me that the order in which the sequence of
> atoms identifying a dihedral or an improper is wrote does not matter, as
> well as the sequence of dihedrals or impropers in aminoacids.rtp file (and
> I agree with you). In any case, in order to try every possible solution, I
> modified the aminoacids.rtp file in order to have exactly the same sequence
> of impropers as in topol.top, and I rewrote all impropers in order to have
> them defined exactly as in topol.top (eg. I converted the CT-CT-N-C
> dihedral into C-N-CT-CT in aminoacids.rtp). I corrected ONLY those parts of
> ffbonded.itp and aminoacids.rtp referred to my chromophore and ONLY in
> dihedrals and impropers, not correcting neither bonds nor angles parameters.
> When I grompp'ed again, I had a new error message, but my surprise was
> great when I found that, instead of the 6 warnings and 17 errors I obtained
> previously, all referred to my chromophore, now I have 2 warnings and 43
> messages, involving not only the dihedrals and impropers of chromophore,
> but also bonds, angles, dihedrals and impropers of other protein residues
> (in particular, two Trp residues). But the really discouraging thing is
> that, when I removed the modified ffbonded.itp and aminoacids.rtp files and
> replaced them with the old ones (that theoretically should cause the old 6
> warnings and 17 errors of which we already discussed), I still have 6
> warnings and 43 errors: replacing these two files does not change the
> errors. I removed all intermediate files and started from the PDB file
> without box, water etc.: nothing changes, I am not able to reproduce the
> same errors I obtained yesterday with the same files (I am sure of it, I
> saved them into dropbox yesterday to render them available and I used
> exactly those files instead of those that I have modified today and that I
> have removed from the amber99sb.ff directory in my home. It sounds like
> Gromacs "remembers" the last modification of ffbonded.itp and/or
> aminoacids.rtp even if these two files have been further modified.

Apparently you are picking up an amber99sb.ff directory that you don't
think you are picking up. That might mean you're starting from a directory
you don't think you are starting in, or referencing some directory other
than the one you mean, or GMXLIB has a different value in a new shell, or
probably some other irritating reasons. You can add "-debug 1" to the
grompp flags to get a grompp.debug output file. The lines in it that begin
with GMXCPP will tell you where grompp is navigating to get the files the
.top file wants (e.g. chdir to note when/where it changes directory, cwd to
report the current working directory, cpp file open when files get opened).
That should help solve part of your issue.

One approach to help save yourself pain is to use a version control or
content management system for your files. Being able to roll back time to
abort a failed attempt is well worth the investment (though perhaps that's
a learning curve for your next project...) Being able to note when progress
has occurred is good for the record (and the psyche!). (And that goes for
everyone else reading this discussion!!)

I am really discouraged about all this story. You say that I'm not doing
> errors, apparently, but things are not going on. The most crazy thing is
> that I DON'T UNDERSTAND where and how I am making mistakes. I don't want
> ask you to do my homework, but I really would like to ask you to check my
> files in order to understand where the error is. I am sure it is somewhere,
> but I am not able to find it.

Well, start with the -debug 1 approach above and see if you're referring to
the files you think you are and that they contain what you think they
contain. The UNIX diff tool can be your friend here - "how do these two
versions of this file differ such that TRP is broken?"

Good luck! :-)


> Anna
> ______________________________**________________
> Anna Marabotti, Ph.D.
> Assistant Professor
> Department of Chemistry and Biology
> University of Salerno
> Via Ponte don Melillo
> 84084 Fisciano (SA)
> Italy
> Phone: +39 089 969583
> Fax: +39 089 969603
> E-mail: amarabotti at unisa.it
> Skype: annam1972
> "When a man with a gun meets a man with a pen, the man with the gun is a
> dead man"
> (Roberto Benigni, about Roberto Saviano)
> Il 10/04/2013 11:30, gmx-users-request at gromacs.org ha scritto:
>> Message: 5
>> Date: Wed, 10 Apr 2013 11:30:05 +0200
>> From: Mark Abraham<mark.j.abraham at gmail.**com <mark.j.abraham at gmail.com>>
>> Subject: Re: [gmx-users] Re: Re: Re: help with a chromophore of GFP
>> To: Discussion list for GROMACS users<gmx-users at gromacs.org>
>> Message-ID:
>>         <CAMNuMASiAQrqZqD_**DWbGYQDg3JO6v0x_**
>> cLDFpXRtZVUnwJouuQ at mail.gmail.**com<CAMNuMASiAQrqZqD_DWbGYQDg3JO6v0x_cLDFpXRtZVUnwJouuQ at mail.gmail.com>
>> >
>> Content-Type: text/plain; charset=ISO-8859-1
>> On Wed, Apr 10, 2013 at 8:49 AM, Anna Marabotti<amarabotti at unisa.it>
>>  wrote:
>>  >Dear gmx-users, dear Mark,
>>> >the soap-opera is going on... and I am so sorry to bother you again and
>>> >again (but I hope that this will be of help for future people that will
>>> >have the same problem...).
>>> >
>>> >Summary of the last episode: I found a way to convert the parameters of
>>> >dihedrals and impropers calculated with Antechamber to Gromacs format,
>>> >and I added these values in ffbonded.itp. Now my ffbonded.itp is
>>> >complete, so I launched grompp on the protein structure with the
>>> >chromophore (I already did pdb2gmx, editconf and genbox with no
>>> >problems), and these are the warnings and the errors I obtained:
>>> >
>>> >grompp_d -f em.mdp -c prot_boxwat.pdb -o prot.tpr -p topol.top
>>> >
>>> >
>>> >WARNING 1 [file ffbonded.itp, line 464]:
>>> >
>>> >   Overriding Proper Dih. parameters.
>>> >
>>> >   Use dihedraltype 4 to allow several multiplicity terms.
>>> >
>>> >   old: 0 8.368 1 0 8.368 1
>>> >
>>> >   new: C   N   CT  CT    9       0.0      8.36800     2
>>> >
>>> >
>>> >
>>> >WARNING 2 [file ffbonded.itp, line 465]:
>>> >
>>> >   Overriding Proper Dih. parameters.
>>> >
>>> >   Use dihedraltype 4 to allow several multiplicity terms.
>>> >
>>> >   old: 0 8.368 2 0 8.368 2
>>> >
>>> >   new: C   N   CT  CT    9       0.0      1.67360     3
>>> >
>>> >
>>> >[other warnings similar to this one, the last warning is:]
>>> >
>>> >
>>> >WARNING 6 [file ffbonded.itp, line 570]:
>>> >
>>> >   Overriding Proper Dih. parameters.
>>> >
>>> >   Use dihedraltype 4 to allow several multiplicity terms.
>>> >
>>> >   old: 0 0 3 0 0 3
>>> >
>>> >   new: H1  CT  CT  OH    9       0.0      1.04600     1
>>> >
>>> >
>>> >Generated 2211 of the 2211 non-bonded parameter combinations
>>> >
>>> >Generating 1-4 interactions: fudge = 0.5
>>> >
>>> >Generated 2211 of the 2211 1-4 parameter combinations
>>> >
>>> >
>>> >
>>> >ERROR 1 [file topol.top, line 31084]:
>>> >
>>> >   No default Improper Dih. types
>>> >
>>> >
>>> >[other 16 error messages similar to these ones]
>>> >
>>> >
>>> >1) Concerning warnings.
>>> >What I understand in the warning is that I defined some values in a
>>> >wrong way.
>>> >In order to understand what was wrong, I looked both at ffbonded.itp and
>>> >to my topol.top file.
>>> >
>>> >In the ffbonded.itp file, all warning messages but one (the last one)
>>> >are referred to some definitions that I added to the file. For example,
>>> >in the ffbonded.itp file, I see the definitions for dihedral CT-CT-N-C,
>>> >that are the followings:
>>> >
>>> >...............
>>> >CT  CT  N   C     9       0.0      8.36800     1  ; new for 99sb
>>> >  CT  CT  N   C     9       0.0      8.36800     2  ; new for 99sb
>>> >  CT  CT  N   C     9       0.0      1.67360     3  ; new for 99sb
>>> >.............................**.**...
>>> >
>>> >Since in the Antechamber parameters I found paramteres for dihedral
>>> >C-N-CT-CT, I decided to add them to the file, in this way:
>>> >...............
>>> >CT  CT  N   C     9       0.0      8.36800     1  ; new for 99sb
>>> >  CT  CT  N   C     9       0.0      8.36800     2  ; new for 99sb
>>> >  CT  CT  N   C     9       0.0      1.67360     3  ; new for 99sb
>>> >C   N   CT  CT    9       0.0      8.36800     1  ; REF. CT-CT-N-C, SET
>>> >  C   N   CT  CT    9       0.0      8.36800     2  ; REF. CT-CT-N-C,
>>> >  C   N   CT  CT    9       0.0      1.67360     3  ; REF. CT-CT-N-C,
>>> >
>> The three lines that I understand you have introduced are not needed.
>> CT-CT-N-C and C-N-CT-CT are the same interaction, and later when grompp
>> looks up the parameters for a [dihedral] from the [dihedraltypes] it looks
>> for either.
>>  >.............................**.**...
>>> >
>>> >The warning message is referred to the second and third parameter I
>>> >added; the first one (marked with "1") is not cited.
>>> >My question is: how to deal with the warning message saying that I am
>>> >"overriding Proper Dih. parameters" and that I should "Use dihedraltype
>>> >4 to allow several multiplicity terms." Do I have to add all 3
>>> >parameters into section "4" instead of "9"? Do I have to remove all the
>>> >three parameters? or only the last two? why the three default parameters
>>> >described just before mine are present as multiple in this section and
>>> >not in section "4"? Moreover, If I understand correctly info in Table
>>> >5.5, dihedraltype 4 should be assigned to "periodic improper dih.", but
>>> >these are not improper dihedral types! How to deal with this problem?
>>> >
>> The error message looks wrong. It was introduced at about the same time as
>> dihedraltype 9, but the history of that part of code development is very
>> murky, and much was not well documented at the time. Even professors are
>> human! I suspect that a subsequent change has broken something that worked
>> well at the time, and because the original purpose was unclear and there
>> were no test cases for the code, nobody knew there was a problem (if one
>> actually exists!). Still, it doesn't matter for your case, because you
>> don't need the duplicate [dihedraltypes] at all.
>> Second question: the last warning message is referred to a parameter
>>> >that is already present in the default ffbonded.itp file. It is defined
>>> >in this way:
>>> >H1  CT  CT  OH    9       0.0      1.04600     1  ; Junmei et al, 1999
>>> >there are no other definitions for this dihedral in this file, and I did
>>> >not add any information about this dihedral (or at least, I did not do
>>> >it intentionally). Why this warning?
>>> >
>> Type 9 is a proper dihedral, per the tables of chapter 5. Presumably you
>> have an entry in an [impropers] section whose atoms match that type. One
>> or
>> other entry needs to move section or change type to get a match.
>>  >2) Concerning the errors.
>>> >I understand that Gromacs does not find correct definitions for
>>> >dihedrals of my chromophore.
>>> >I don't understand if this comes from the definition of dihedrals that I
>>> >included in aminoacids.rtp or from the parameters that I included in
>>> >ffbonded.itp.
>>> >
>>> >In aminoacids.rtp I added the section [impropers] just below the section
>>> >[bonds]. Impropers are defined like that, using the atom names (that are
>>> >the same as in the .pdb file that I grompp'ed):
>>> >  [ impropers ]
>>> >     -C   CA1     N   H04
>>> >    CA4    +N     C     O
>>> >     C2   CB2     N   CA2
>>> >.............................**.**..................
>>> >    CZ1   CD2   H14   CE2
>>> >    CE2   CG1   H10   CD2
>>> >When I looked at the lines in topol.top cited in the error message, I
>>> >found that for several impropers, the definition was the opposite with
>>> >respect to the one that I included in aminoacids.rtp. For example, in
>>> >topol.top the atoms composing the last improper of aminoacids.rtp
>>> >(CE2-CG1-H10-CD2) are indicated exactly in the opposite order
>>> >(CD2-H10-CG1-CE2). Instead, in other impropers, atoms are reported
>>> >exactly in the same order in both files. Moreover, the order in which
>>> >impropers are present in the aminoacids.rtp file is not the same that I
>>> >find in topol.top (eg. the last improper CE2-CG1-H10-CD2 in
>>> >aminoacids.rtp is not the last improper in topol.top). Do these two
>>> >differences matter for the correct recognition of dihedrals?
>>> >
>> As above, a dihedral interaction is defined over an ordered set of 4
>> atoms,
>> but reversing the order describes the same interaction. grompp doesn't
>> bother with any aesthetics here; even if it wanted to, and if the .rtp
>> entry and the [dihedrals]/[impropers] section were in the opposite sense,
>> there's no clearly correct "orientation" to use. Similarly for interaction
>> ordering. The one-to-one mapping in the lookup is important, and nothing
>> else is.
>> Second question: I decided not to implement all improper parameters
>>> >calculated with Antechamber on my molecule. In fact, they were
>>> >calculated using GAFF instead of Amber99sb FF (the two ff are fully
>>> >compatible, but since some of these parameters are referred to moieties
>>> >of the chromophore that are easily recognizable as former amino acid
>>> >residues, I prefer to use the classical Amber99sb parameters for these
>>> >residues).
>> Seems sensible, but do be aware that Antechamber has optimized the other
>> parameters for a context other than that in which you plan to use them.
>> Whether that is significant is difficult to show. What is easy to do (and
>> I
>> think I have said this before in this thread) is to check the docs to see
>> if you can do the original Antechamber calc in the constrained context. If
>> not, do hassle the Antechamber developers to document/implement how to do
>> it! You're not the only one who would want this functionality.
>> Therefore, I did not add for some of these impropers the
>>> >specific parameters, preferring to use the general ones already defined
>>> >in ffbonded.itp. Can this fact contribute for this error?
>>> >
>> The warnings here pertain to possible violations of the assumptions by
>> which grompp tries to look up parameters for intractions in your
>> [moleculetype]. The errors occur when grompp cannot do that look-up.
>> Omitting what looks like a duplicate would lead to an error if you aren't
>> as picky as the computer program has to be in doing the look up. I think
>> your procedure looks fine, except for the proper/improper mismatch above.
>> Mark
> --
> 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