[gmx-users] Problem with encad in pdb2gmx, GMX 3.3 CVS version. Posible Bug in ffencads-c.tdb file?

[Leonardo Sepulveda Durán]

Hello!!!

A few weeks ago I posted an error I got using pdb2gmx and encads :

pdb2gmx  -f S1.pdb -ignh

OUTPUT:

Number of bonds was 1059, now 1059
Generating angles, dihedrals and pairs...
Before cleaning: 2808 pairs
-------------------------------------------------------
Program pdb2gmx, VERSION 3.3
Source code file: pgutil.c, line: 87

Fatal error:
Atom CB not found in residue 64 while adding improper

-------------------------------------------------------

The file numbering begins with 20, and end in 83, so the trouble is in
the last residue. I checked the presence of two oxigen atoms, but
leaving only one do not solve the trouble.

whwn I use other protein the program works. Then I cut that protein on
a GLY residue, giving this terminal atoms

ATOM    315  N   GLY    39       5.478  39.842  35.478  1.00 13.50      1ENH 376
ATOM    316  CA  GLY    39       6.343  40.900  35.994  1.00 12.21      1ENH 377
ATOM    317  C   GLY    39       5.897  41.429  37.359  1.00 12.24      1ENH 378
ATOM    318  O   GLY    39       6.219  42.547  37.738  1.00 11.54      1ENH 379
TER     419         GLY    39                                         
    1ENH 528
END

The error appear again. I think this would be a bug regarding the
patching of terminal glycines using encad (this dont happen using
OPLS). I tried to track the error, but pgutil.c only have subroutines
for printing.   Looking at ffencads-c.tdb this appears

[ COO- ]
[ replace ]
C               Abis    12.011  0.84
[ add ]
2       8       O       C       CA      N
        Obis    15.9994 -0.92
[ delete ]
O
[ dihedrals ]
N       CA      C       O2
[ impropers ]
CA      O2      C       O1      improper_X_X_A_X
 N       C       CA      CB      improper_X_X_C_X

It seems the last line is the problem, the program try to delete that
improper but it cant find a CB atom in Glycines. The OPLS file for
patches (ffoplsaa-c.tdb), is longer, it have special treatments for
GLY and PRO.

I delete the last line (N       C       CA      CB     
improper_X_X_C_X) and the problem was solved. Nevertheless, this would
be detrimental for other residues with CB atoms, and I really dont
know if my fix would give another trouble. Can someone help? Erik?
David?

Leonardo