Hi,
I have a question concerning the value of the sasopt option when using
pbsa for PB analysis within MMPBSA.py in the case one uses inp=1. If I
understand correctly, inp=1 implies use_sav=0. In this case there is no
dispersive component in the non-polar part, and the cavity energy is
calculated from the area of the solvent accessible surface (SAS) with
cavity= a*SASA + b.
Now, I see in the manual that pbsa uses sasopt to control the type of
molecular surface that is calculated. The default is saspot=0 wich
calculates the solvent excluded surface (SES). SES is sensibly smaller
than SAS, so I thought that using inp=1 I should use sasopt=1 to require
SAS calculation.
However if I force sasopt=1 I get one of the famous PB Bombs, for example:
(...)
> ======== FDPB Summary ========
>
> Do FDPB every 1 steps
> Nonbonded Update
> residue cutoff is set to 99.0000000000000
> fdpb cutoff is set to 16.0000000000000
> sas cutoff is set to 16.0000000000000
> nonbonded cutoff is set to 99.0000000000000
> Grid Constants
> Grid dimension 65 59 65
> Grid spacing set to 4.000
> Grid boundary
> -129.000 135.000
> -121.000 119.000
> -125.500 138.500
>
> Dielectric Map
> Use cavity radii in the prmtop file
>
> Use level-set-based SES definition
> Compute SAS/SAR every 1 steps
> Solvent probe radius 1.40000000000000
> Surface dots per atom 366
>
> Boundary conditions
> sum of grid charges as independent DH spheres
>
> Physical constants
> Solute dielectric constant : 1.00000000000000
> Solvent dielectric constant : 80.0000000000000
> Temperature (K) : 300.000000000000
> Ionic strength (mM) : 150.00000000000000
> Debye-Huckel parameter (1/A): 0.12572388225794
>
> FD Solver Option
> Use Modified ICCG solver
>
> Iteration data
> Maximum iterations : 50000
> Convergence criteria: 1.000000000000000E-06
> Iterations required : 7
> Norm of the constant vector: 23.1639921089
> Norm of the residual vector: 0.0000026901717
> Convergence achieved : 1.1613592503166168E-07
>
> PB Bomb in assignlvlset(): no atmsas 70 53 6
> 72 55 8 -2
> 72 55 8 -2
> 72 55 8 -2
> 72 55 8 -2
> 72 55 8 -2
> 72 55 8 2
> 72 55 8 -2
> PB Bomb in assignlvlset(): illegal insas flag 70 53 6 -4
If I don't set sasopt, the calculations are performed with no error.
This is the input for MMPBSA.py (by the way, if someone sees anything
wrong/strange I would appreciate your comments):
> Input file for running mmpbsa using PBSA
> &general
> debug_printlevel=1,
> endframe=1000000000,
> entropy=0,
> interval=5,
> keep_files=1,
> ligand_mask=':70-141',
> netcdf=1,
> receptor_mask=':1-69,142-183',
> search_path=1,
> startframe=1,
> strip_mask=':WAT,Cl-,CIO,Cs+,IB,K+,Li+,MG2,Na+,Rb+',
> use_sander=1,
> verbose=1,
> /
> &pb
> # for inp=1
> inp=1,
> cavity_offset=0.86,
> cavity_surften=0.00542,
> radiopt=0,
> # for inp=2, which cannot be used with ff10 atom types...
> # inp=2,
> # cavity_offset=-0.5692,
> # cavity_surften=0.0378,
> # radiopt=1,
> exdi=80.0,
> fillratio=4.0,
> indi=1.0,
> istrng=0.15,
> linit=50000,
> prbrad=1.4,
> sander_apbs=0,
> scale=2.0,
> /
To introduce the sasopt that I want and also because I want to use the
non-linear solver (the system includes a DNA), I hack the mdin files
produced by MMPBSA.py so they finally look like:
> File generated by MMPBSA.py
> &cntrl
> nsnb=99999, dec_verbose=0, ioutfm=1,
> ipb=2, ntb=0, cut=999.0, imin=5,
> inp=1, igb=10,
> /
> &pb
> use_sav=0, sasopt=1, sprob=1.4, dprob=1.4,
> npbopt=1, eneopt=1, bcopt=5, cutnb=99.0, cutfd=16.0,
> npbverb=1, accept=0.000001,
> istrng=150.0, cavity_offset=0.86,
> maxitn=50000, fillratio=4.0,
> radiopt=0, cavity_surften=0.00542,
> /
BTW, the reason I am using inp=1 is that the prmtop files were prepared
with LEAP using ff10 (to have ff99bsc0 for the DNA) and this includes
atomtypes not supported by pbsa with the inp=2 option.
Is my interpretation about the use of the sasopt option correct?
Any clues about what may be producing the error?
Cheers,
--
Miguel Ortiz Lombardía
Architecture et Fonction des Macromolécules Biologiques (UMR7257)
CNRS, Aix-Marseille Université
Case 932, 163 Avenue de Luminy, 13288 Marseille cedex 9, France
Tel: +33(0) 491 82 55 93
Fax: +33(0) 491 26 67 20
mailto:miguel.ortiz-lombardia.afmb.univ-mrs.fr
http://www.afmb.univ-mrs.fr/Miguel-Ortiz-Lombardia
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Wed Feb 13 2013 - 06:00:06 PST
