[gmx-users] Using specbond.dat in pdb2gmx

[Justin Lemkul]

On 6/13/14, 12:57 PM, Matthew Stancea wrote:
> Dr. Justin Lemkul,
>
> Now I understand the message about removal of bonds being just duplicate bonds that have been removed, and, also, I understand how the -ter flag works now.
>
>>> This is the screen output related to parsing of specbond.dat:
>>>
>>> "3 out of 3 lines of specbond.dat converted successfully
>>> Special Atom Distance matrix:
>>>                       CYS1    CYS1    CYS5    CYS5    VAL6   CYS10   CYS10
>>>                         N1     SG6     N47    SG52     C59    N101   SG106
>>>       CYS1     SG6   0.302
>>>       CYS5     N47   0.988   0.900
>>>       CYS5    SG52   1.070   0.917   0.333
>>>       VAL6     C59   1.498   1.396   0.510   0.585
>>>      CYS10    N101   1.122   0.901   0.614   0.609   0.835
>>>      CYS10   SG106   0.879   0.655   0.490   0.422   0.857   0.322
>>>      CYS15    N160   0.784   0.488   1.066   0.984   1.477   0.774   0.620
>>>      CYS15   SG165   0.477   0.202   0.960   0.961   1.432   0.823   0.625
>>>      CYS17    N184   1.030   0.784   0.721   0.460   1.004   0.617   0.387
>>>      CYS17   SG189   1.224   1.039   0.514   0.204   0.625   0.621   0.487
>>>      VAL21    C245   0.756   0.685   0.408   0.388   0.847   0.812   0.531
>>>      CYS22    N259   0.657   0.574   0.404   0.432   0.887   0.742   0.458
>>>      CYS22   SG264   0.755   0.588   0.371   0.430   0.827   0.436   0.203
>>>      VAL29    C363   0.132   0.394   1.117   1.193   1.626   1.248   1.004
>
>> Here is precisely why your specbond.dat is not working - the bond criteria are
>> not satisfied.  The Cys1(N)-Val29(C) distance is 0.132 nm.  You're specifing in
>> specbond.dat (likely from simply copying the contents of the other lines) that
>> the reference distance is 0.25 nm.  If Gromacs does not find atoms within ±10%
>> of the value in specbond.dat, a bond won't be created.  Since 0.132 nm is way
>> off, you won't get a bond.
>
>> Also note that your final columns will try to rename the residues to CYS2 (which
>> is specific for a disulfide cysteine, so unless that's true you'll get more
>> fatal errors) and VAL2, which doesn't exist.
>
>> See http://www.gromacs.org/Documentation/File_Formats/specbond.dat
>
>
> Ah I see. I also noticed that in the original specbond.dat, the bond distance specification for the disulfide bonds between 2 CYS's or between 2 CYM's are .2 nm, not .25 nm, so I changed that as well. The only thing I still find a bit unclear is the final columns (residue rename). Is this what my specbond.dat file should look like?
>

The final columns specify the new residue names that should be assigned to the 
residues only in the case that a special bond was created.  You'll see from the 
existing Cys specifications that a residue named "CYS" (standard PDB 
nomenclature for any Cys residue) is converted into the .rtp-specific CYS2, but 
only if a special bond is created.  The nomenclature exists because of the way 
the force fields work.  In your case, you don't need to rename anything, so 
preserving the original residue names is the proper way to go.

Of course, the answer to "is this correct?" is generally obtained by running it ;)

> 3
> CYS     N       1       VAL     C       1       0.132    CYS    VAL
> CYS     SG      1       CYS     SG      1       0.20    CYS2    CYS2
> CYM     SG      1       CYM     SG      1       0.20    CYS2    CYS2
>
> -Matthew Stancea
>
> (This email chain has gotten a bit long in my opinion, so I have put the previous email below. With your permission, I would prefer just using the above message in order to continue correspondence in shorter emails; however, if you would prefer to instead use the entire email, I completely understand.)
>

Snipping out relevant portions is perfectly fine (and preferred) as long as you 
don't nuke something actually important.

-Justin

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

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

Department of Pharmaceutical Sciences
School of Pharmacy
Health Sciences Facility II, Room 601
University of Maryland, Baltimore
20 Penn St.
Baltimore, MD 21201

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

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