Tutorial on using Qnifft
I.
What you need
II.
The Qnifft parameter file
III.
Common problems
I. What you need
- First, you need to have the Qnifft program. If you need to download
the program, it can be found here.
This program can be run only on a UNIX machine, such as an SGI O2.
- Second, you need to have a PDB file of the structure of interest. The
atoms must be labeled in Insight II style.
- Fourth, you need to have parameter files for both atomic radii and
partial atomic charges. Examples of these files are here.
These files can be modified to any specifications that you wish.
- Fifth, you need Grasp,
or a related program, so that you can view the results of the calculations.
II. The Qnifft
parameter file
This section contains information on the parameters that can be altered
during an electrostatics calculation run. The sample file is called prm.
There are instructions on how to alter the parameters with the prm file.
- Gridsize
The largest gridsize available with this version of Qnifft is 65x65x65.
Smaller gridsizes can be used but larger ones cannot (default=65).
- Scaling
You must choose one of the following to determine how the molecule fits
within the grid:
- Boundary Conditions
There are four choices for boundary conditions (I do not know what the
"field" condition does). The condition must be inputted as boundary_condition=xxx
- Zero: All boundary points have zero potential. This is a rather
inaccurate method but requires minimal computing. Most useful if the molecule
being analyzed is very small.
- Dipolar: All boundary points have coulombic potentials as calculated
from the center of positive charge and from the center of negative charge.
This method is more accurate than "zero" and also requires minimal
computing.
- Coulombic: All boundary points have coulombic potentials as
calculated from each point charge within the system studied. This is much
more accurate than "zero" or "dipolar" but requires
more computing (default setting).
- Focussing: This requires two calculation runs. The first run
should be done with one of the other conditions above, but at low resolution
(i.e. at low grids per angstrom). The phi-map generated in the first run
is then used as the basis for calculating the boundary potentials in the
second "high-resolution" run.
- Dielectric: Input the value for the dielectric of the interior
of the molecule (default=2).
- Spherical charge: Usually kept as false (default=f).
- Dielectric: Input the value for the dielectric to the exterior
of the molecule (default=80).
- Salt concentration: Input the value (in molar) for the concentration
of a 1:1 salt to be calculated within the solvent (default=0.145,
which gives a Debye-Huckel radius of 8 angstroms).
- Probe radius: Input the radius size of a sphere which will "roll"
on the surface of the molecule. This determines the accessible van der
Waals surface (Conolly method???) of the molecule (default=1.4).
- Ionic radius: Input the radius of a sphere which will roll on
the surface of the molecule. The distance from the center of this sphere
to the van der Waals surface of the molecule determines an area of solvent
that is inaccessible to ions in the solvent (i.e. dielectric=80, ionic
strength=0) (default=2.0).
- Temperature: Input the temperature (in Kelvin) for which the
calculations shall be done (default=298).
- Nonlinear equation: Determines if the nonlinear Poisson-Boltzmann
equation is to be used in calculating electrostatics (default=t).
- Convergence: Determines the value at which the iterations are
considered to be converged (default=0.001).
- Relaxation parameter: Kept as 1.0 (default=1.0).
- Check frequency: Kept as 2 (default=2).
- Newton iterations: Determines maximum number of iterations to
be performed (default=20).
- Level 0 multi-grid: Kept as 2 (default=2).
- Level 1 multi-grid: Kept as 4 (default=4).
- Level 2 multi-grid: Kept as 8 (default=8).
- Level 3 multi-grid: Kept as 16 (default=16).
- Smooth dielectric: Kept as 0 (default=0).
- Concentration output: ? (default=f)
- Insight format: Determines if output pdb files will be in Insight
format (default=f).
- Site potentials: Determines if a file containing site files
will be output (default=f).
- Atom file output: Determines if a file containing atoms and
their assigned partial charges will be output (default=f).
- Analyse map: ? (default=f)
- Title: ? (default=untitled)
- PDB input file: Specifies a PDB file on which the electrostatic
calculations will be performed.
- Phi input file: Specifies a phi-map to be used in calculating
boundary potentials. This is used in the focussing technique.
- Site input file: Specifies a PDB file whose site potentials
will be determined.
- PDB output file: Gives a name to a file to be written that contains
the PDB file used as the partial charges assigned to atoms during the calculation.
- Phi ouput file: Gives a name to a file to be written that contains
the phi-map (Grasp readable) for the calculations performed.
- Dielectric map file: Gives a name to a file that shows the dielectric
map.
- Site output file: Gives a name to a file to be written that
contains electrostatic potentials at sites provided by the site input file.
III. Common
problems
- could not find radius/charge for atom specified: probably missing
an atom name; check that atoms are appropriately
named to fit with the nomenclature of the radius/charge files
- parameter missing: either a necessary parameter (i.e. the PDB input
file, dielectrics, etc.) or a optional parameter specifically turned is
missing; see which parameter needs a value and modify
the parameter file accordingly
- iterations do not converge or do not reach the convergence value:
the system is probably too large; increase the salt
concentration or try to decrease the size of the system; or calculations
have reached the maximum iterations; increase the
number of iterations
- shortest distance from molecule to boundary is a negative number:
scale is too large; decrease the scale or use the
fill or border scaling options
- salt points outside the grid: scale is too small; increase
the scale (I have seen this during the first run of a focussing calculation.
However, I'm not sure if this causes any harm for a focussing calculation.)
- If there are further problems or questions, please contact
me.
|