Martini RBCG

From: DENILSON FERREIRA DE OLIVEIRA (denilson_at_dqi.ufla.br)
Date: Mon Jan 14 2013 - 09:55:51 CST

Dear all,

In order prepare my protein to a Martini RBCG Molecular dynamics simulation, I used the information and files (topology, parameter, scripts etc.) of the Residue-Based Coarse-Graining tutorial (http://www.ks.uiuc.edu/Training/Tutorials/martini/rbcg-tutorial.pdf and http://www.ks.uiuc.edu/Training/Tutorials/martini/files.tar.gz). At first, I created the coarse-grained model of my protein using the martini-protein.cgc file. No error (or warning) message could be found in the log file. In the generated pdb file I could see the following lines around atoms 86 and 87:

ATOM 85 SI2 TYR P 34 83.520 87.398 184.728 1.00 0.00 P1
ATOM 86 BAS GLY P 35 77.848 84.404 184.186 1.00 0.00 P1
ATOM 87 BAS ALA P 36 77.655 81.077 183.112 1.00 0.00 P1
ATOM 88 BAS GLN P 37 77.412 81.465 179.915 1.00 0.00 P1

Then, I used the file martini-protein.top to generate a new pdb file and the psf file. Again, no error (or warning) message could be found in the log file. Around atoms 86 and 87 I could see the following lines in the generated psf file:

85 P1 34 TYR SI3 SP1Y 0.000000 45.0000 0
86 P1 35 GLY BAS P5 0.000000 72.0000 0
87 P1 36 ALA BAS P4 0.000000 72.0000 0
88 P1 37 GLN BAS P5 0.000000 72.0000 0

To correct this psf file, I used the fix_martini_psf.tcl script with dssp (the same results were obtained with stride). No error (or warning) message could be found in the log file generated by the script or dssp. The region around atoms 86 and 97 looked like the following lines in the new psf file:

85 P1 34 TYR SI3 SP1Y 0.000000 45.0000 0
86 P1 35 GLY BAS P5C 0.000000 72.0000 0
87 P1 36 ALA BAS P4 0.000000 72.0000 0
88 P1 37 GLN BAS N0H 0.000000 72.0000 0

I solvated and ionized my protein using the files in the tutorial to obtain the final coarse-grained model of my protein in a water box. The lines around atoms 86 and 87 were not modified:

85 P1 34 TYR SI3 SP1Y 0.000000 45.0000 0
86 P1 35 GLY BAS P5C 0.000000 72.0000 0
87 P1 36 ALA BAS P4 0.000000 72.0000 0
88 P1 37 GLN BAS N0H 0.000000 72.0000 0

Although everything seemed to be correct, when I tried to minimize the structure using namd 2.8 and the parameter files of the tutorial (martini-all-nonb.par, martini-lipids-bonds-angles-dihedrals.par, martini-protein-angles-cos.par, martini-protein-bonds.par and martini-protein-dihedrals.par), namd could not find some bond parameters:

FATAL ERROR: UNABLE TO FIND BOND PARAMETERS FOR P4 P5C (ATOMS 86 87)

In fact, I could not find bond parameters for P4 in the file martini-protein-bonds.par. I could also find in the same parameter file that:

!ALA (none)
!ALA can have BAS = P4C, P4B, P4F, C5H, N0H, N0T, N0E
! or BAS = QdC, QaC, QdB, QaB, QdH, QaH, QdT, QaT, QdEX, QaEX for charged NTER/CTER
! or BAS = NaC, P3C, NaB, P3B, NaT, P3T, NaEX, P3EX, NaH, P3H for uncharged NTER/CTER
!but ALA has no sidechain in this model, so this part is empty.

Consequently, ALA could not have BAS = P4. It should have BAS = P4C, P4B, P4F, C5H, N0H, N0T or N0E. However, as ALA (atom 87) was considered POLAR (P4), the only alternatives in this case for the atom 87 are P4C, P4B and P4F. Therefore, I walked through the log file of the script fix_martini_psf.tcl and found the following lines:

Secondary structure is C. Applying patch OCC P1:34 to segname P1 resid 34.
Secondary structure is C. Applying patch GCC P1:35 to segname P1 resid 35.
Secondary structure is H. Applying patch ACN P1:36 to segname P1 resid 36.
Secondary structure is H. Applying patch OCH P1:37 to segname P1 resid 37.

Thus, according to the stride format (used by the script fix_martini_psf.tcl) atom 86 (resid 35) is in a coil (C) and atom 87 (resid 36)is in a alpha helix (H). Consequently, BAS = P5C for atom 86 (resid 35) seems to be correct, but atom 87 (resid 36) should be P4H instead of P4. However, P4H is not available for ALA in the martini-protein-bonds.par. As P4B apparently corresponds to isolated bridge (B), it seems that BAS for atom 87 (resid 36) must be P4F. Correct ?

I would appreciate if someone could be kind enough to help me.

Best regards.

Denilson.

Denilson F. OliveiraLaboratório de Produtos NaturaisDepartamento de QuímicaUniversidade Federal de LavrasCaixa Postal 3037Lavras - MG - BrasilCEP 37.200-000Tel: (55)(35) 3829-1623Fax: (55)(35) 3829-1271e-mail: denilson_at_dqi.ufla.br

This archive was generated by hypermail 2.1.6 : Tue Dec 31 2013 - 23:22:53 CST