Opcodes for Selection of Force Field, Nonbonded Cutoffs, and Solvation Treatment

The opcodes in this section are linked below:

FFLD — Force FieLD Selection

FFOP — Force Field OPtion selection

EXNB — use EXtended noNBonded cutoffs

EXN2 — EXteNds EXNB

BDCO — Bond Dipole CutOffs

SOLV — SOLVation selection

LOGP — LOG P partition coefficient estimation

CHGF — CHarGe File

FFLD — Force FieLD Selection

Specifies a force field; there is no default for this command. A force field must always be specified.

The force-field files we supply have names fieldname.fld, where fieldname is one of: mm2, mm3, amber, amber94, oplsa, f10 or f14. MacroModel looks in the following directories in the order listed for the specified force-field file and uses the first one it finds:

  • The local directory
  • The macromodel subdirectory of your Schrödinger user resources directory.
  • The product data directory ($SCHRODINGER/macromodel-vversion/data).
  • The mmshare data directory ($SCHRODINGER/mmshare-vversion/data).

Before looking for fieldname.fld, MacroModel looks locally for a file called filename.fieldname (e.g., my_job.amber). Under certain circumstances, other processes preparing jobs for running by MacroModel use this convention.

Parameters for more recent force fields (arg1 value of 10 or higher) are obtained from a coprocess interacting with MacroModel using the BMFF mechanism. See The BMFF Protocol for details. The force-field file name used in this case is f10.fld for MMFF and f14.fld for OPLS_2005.

 

arg1

Force field

1

MM2*. Allinger’s 1987 parameter set with many additions [1, 2]. Used for simple organics. Differs from the authentic field by use of a Coulomb’s law treatment of electrostatics and torsional barrier treatment of conjugation.

2

MM3*. Allinger’s 1990 parameter set with additions [3]. Used for simple organics. Differs from the authentic field by use of a Coulomb’s law treatment of electrostatics and torsional barrier treatment of conjugation.

3

AMBER*. Kollman’s united atom and all atom fields with additional parameters for organic functionality [4, 5]. Used primarily for biopolymers. The MacroModel default for hydrogen bonding uses Kollman’s 6,12-Lennard Jones treatment [6] and an improved peptide backbone parameter set [7].

4

AMBER94. Kollman’s 1994 version of AMBER, Amber4.1 [8].

5

OPLSA*. Jorgensen’s nonbonded parameter set + AMBER bonded functions for liquid simulations [9]. Used primarily for peptides. Best for relatively rigid molecules, because the torsional parameters have not been optimized to reproduce conformational energy differences.

10

MMFF94 and MMFF94s [10–16]. See also the description of arg4 of this command.

14

OPLS_2005. An enhanced version of the original OPLS_2001 all-atom force field, which was developed by Professor W. Jorgenson of Yale University and was at that time probably the best one available by default in MacroModel for condensed-phase simulations of peptides [17]. OPLS_2005 was developed by Schrödinger to provide a larger coverage of organic functionality. In particular all torsional parameters have been refit to reproduce the conformational energetics derived at a higher level of quantum theory and additional charges have been fit to support additional organic functionality. The parameters for proteins have been updated to the ones published more recently [53].

16

OPLS4. The latest improvement in the OPLS-AA force fields, extended to significantly improve accuracy in many areas.

arg2

Electrostatic treatment

0

Default. Uses dielectric treatment encoded within force field file unless solvation model 3 is used (see SOLV command), in which case the constant dielectric treatment is used.

Note: All our force fields are supplied, by default, with constant dielectric electrostatics. Prior to MacroModel 6.0, AMBER*, MM2, and MM3 used distance-dependent dielectric constant by default.

−1

Turns Coulombic molecular electrostatics off.

1

Gives constant dielectric electrostatics.

2

Gives distance-dependent dielectric electrostatics.

arg3

Hydrogen bonding treatment

0

Uses equation selected in force field file (default).

1

Turns off explicit 10,12 hydrogen bonding function and uses 6,12 Lennard Jones instead.

2

Gives explicit 10,12 hydrogen bonding function.

arg4

BMFF force-field option
 

For a BMFF force-field, this requests a special option given in an “Option:” line of the MacroModel force-field file. For MMFF, a “1” in this position requests that MMFFs parameters be used; these enforce planarity about delocalized sp2 nitrogens.

arg5

Molecular dielectric constant
 

Default = 1.0. The solvent dielectric constant is normally read from the solvent file (see SOLV command) and should not generally be set here.

arg7

Planarity of aromatic rings
 

Enforce planarity of aromatic rings for OPLS_200X force fields.

0

Do not enforce planarity of aromatic rings. This is the default.

1

Enforce planarity of aromatic rings.

arg8

Coefficient for torsional damping
 

Default: 100.
 

For highly strained conformations in which bond angles become straight, related dihedral angles can become undefined. To avoid this problem and the physically rapidly varying dihedral angle potentials for states close to this, the torsional potential is damped using the function:
 

fdamp = (1 − exp(−ctors(1−cos2θ1))) (1 − exp−ctors(1−cos2θ2)))
 

where θ1 and θ2 are the bond angles formed by the first and last three atoms in the torsional potentials.
 

The default value of the torsional damping coefficient, ctors, of 100 damps the dihedral angle potential significantly for bond angles less than 10 degrees and works well. This argument permits one to change the value of ctors. Increasing ctors dramatically reduces the range over which the damping is applied.

FFOP — Force Field OPtion selection

Specifies alternative (ALT) selections that override those in the force field file. May be used, for example, to select different Z0 atom definitions. See Force-Field File Format for details. An FFOP command must come before an FFLD command

arg1

ALT number
 

The alternative being selected in the force field file.

arg2

ALT selection
 

Number corresponding to the ASCII character defining the selection desired for the alternative given in arg1. The numbers corresponding to the various ASCII characters are given in the table below.

 

Table 1. ASCII to decimal number conversion table.

Char

Num

Char

Num

Char

Num

Char

Num

Char

Num

)

 41

=

 61

Q

 81

e

101

y

121

*

 42

>

 62

R

 82

f

102

z

122

+

 43

?

 63

S

 83

g

103

{

123

,

 44

@

 64

T

 84

h

104

|

124

-

 45

A

 65

U

 85

i

105

}

125

.

 46

B

 66

V

 86

j

106

~

126

/

 47

C

 67

W

 87

k

107

 

 

0

 48

D

 68

X

 88

l

108

 

 

1

 49

E

 69

Y

 89

m

109

 

 

2

 50

F

 70

Z

 90

n

110

 

 

3

 51

G

 71

[

 91

o

111

 

 

4

 52

H

 72

\

 92

p

112

 

 

5

 53

I

 73

]

 93

q

113

 

 

6

 54

J

 74

^

 94

r

114

 

 

7

 55

K

 75

_

 95

s

115

 

 

8

 56

L

 76

`

 96

t

116

 

 

9

 57

M

 77

a

 97

u

117

 

 

:

 58

N

 78

b

 98

v

118

 

 

;

 59

O

 79

c

 99

w

119

 

 

<

 60

P

 80

d

100

x

120

 

 

EXNB — use EXtended noNBonded cutoffs

Despite its name, this command can be used to specify short as well as long cutoffs.

Extended cutoff distances are, by default, 8 Å in van der Waals, 20 Å in charge/charge electrostatics. Standard defaults in the absence of this command are 7 Å for van der Waals and 12 Å for charge/charge electrostatics. Other cutoffs may be selected by adding values for arg5-8. Calculations dealing with ions should use the EXNB option.

Large distance values for cutoffs generally slow calculations but often make convergence smoother. Occasional problems with energies and gradients which appear to increase upon repeated minimizations may usually be solved by using long van der Waals and electrostatic cutoff distances. The native MM2/MM3 and MMFF programs use complete pair lists (no cutoffs) for van der Waals and electrostatic interactions.

arg 1

Long-range derivative update interval
 

Iterations (timesteps) per recalculation of long range (>5 Å) nonbonded derivatives (Default: 10), except when one of the following special values is used:

1

The constant long-range derivative option is turned off, and the entire pair list is used in the evaluation of nonbonded derivatives.

2

All nonbonded pairs are put on the pair list, and the pair list is never updated. Long-range derivatives are used.

3

All nonbonded pairs are put on the pair list, the pair list is never updated, and the entire pair list is used in each evaluation of nonbonded derivatives.

arg 2

Long-range derivative distance cutoff
 

Distance (integer, in angstroms) for distinction between close- and long-range nonbonded interactions, used in constant derivative option. Default value is 5. Longer distances give more accurate derivatives but slow the calculation.

arg5

van der Waals cutoff

arg6

Coulombic electrostatic cutoff

arg7

Hydrogen bonding cutoff
 

Default: 4.0 Å. There is usually no reason to change this.

arg8

Fmm cutoff
 

Any fixed or frozen atoms are placed within this distance of a moving (SUBS) atom in a special class called Fmm. The Fmm atoms are treated in greater detail than the other fixed or frozen atoms in GB computations. All fixed and frozen atoms have their GB radii properly recomputed when the moving atoms move. When the polarization free-energy (Gpol) is computed, fixed-moving interactions are computed using these updated GB radii. However, the GB contributions from pairs of fixed or frozen atoms utilize the updated GB radii only when both fixed atoms are Fmm. Tests show that a reasonable value for CutFmm is 8.0 Å, which is the default.
 

Regardless of the setting of arg8, the program does not allow the Fmm cutoff to exceed the larger of the van der Waals and the electrostatic cutoff.

0

Default: 8.0 Å

> 0

Interpreted as cutoff in Å.

< 0

Interpreted as 0.

EXN2 — EXteNds EXNB

This sets CutsFm, the assumed maximum distance in angstroms that two atoms can be from each other and still influence each others’ solvent-exposed surface areas. This cutoff is used for one-time computations of solvent-exposed surfaces with fixed or frozen atoms.

arg 5

CutsFm cutoff

0

8.0 Angstroms (default).

nn Angstroms.

BDCO — Bond Dipole CutOffs

This option turns on the use of Bond Dipole Cutoffs for the truncation of electrostatic and GB interactions. Two parameters can be specified: the cutoff distance for charge-dipole interactions and the cutoff distance for charge-charge interactions. The cutoff distance for dipole-dipole interactions is taken to be EXNB arg6. If EXNB is not specified, the default value for electrostatic cutoff distance is used (see the EXNB opcode description).

Limitations:

  • If the EXNB opcode is used in the .com file, it must precede the BDCO opcode.

  • Use of the BDCO opcode automatically turns on DEBG flags 89 and 90. BDCO treats all atoms, including hydrogens, explicitly for the purposes of generalized Born solvation. However, if the file structure itself employs united atoms, such atoms are treated as united atoms. In other words, all atoms in the structure file are treated as they are for the purposes of generalized Born solvation in conjunction with BDCO.

    arg5

    Cutoff distance for charge-dipole interactions

The default value is sqrt(cutes**3) where cutes is the electrostatic cutoff distance as per EXNB.

arg6

Cutoff distance for charge-charge interactions

The default value is 9999999.0, which effectively includes all such interactions.

These default settings typically yield shorter nonbonded pairlists than using residue-based cutoffs and the same value for EXNB arg6.

Related DEBG flags: 121, 126, 127, 128, 129.

SOLV — SOLVation selection

Specifies a solvation model (arg1) and a solvent (arg2), so that energy calculations include the approximate effects of solvent.

Solvation model 1 involved explicit solvent and is no longer supported.

Models 2 and 3 read the appropriate solvent file named solvent_name.slv. The file water.slv is available for models 2 and 3, while chcl3.slv and octanol.slv are available for model 3. MacroModel uses the same search path to locate the .slv file as for the force field—see FFLD — Force FieLD Selection.

Solvent model 2 (arg1 = 2) is purely a surface-area-based model. We recommend model 3 for all computations where solvation energies are desired. Model 2 operates as described by Hasel et al. [18]. See also Ooi et al. [20] for the parameter set given in water.slv. Solvent model 3 provides a volume-based continuum model (the GB/SA model) for the electrostatic (polarization) component. [21]

Using model 3, molecular electrostatics should be carried out with a constant dielectric treatment and a low molecular dielectric constant (e.g., 1.0). Constant dielectric electrostatics will be set automatically whenever solvent model 3 is used regardless of the default electrostatic equation selection in the force field file. EXNB should also be used with solvent model 3.

Parametrizations specific for OPLS-AA and MMFF(s) have been added to the water.slv file for model 3. Other force fields continue to use the default parametrization which is unmodified from previous releases. The default parametrization in the octanol.slv file was constructed for the MMFF force field. As well, a parametrization specific for OPLS-AA is present in the octanol.slv file. The new parameters for MMFF are based on the parametrization described in [22] for water, and in [23] for octanol. We gratefully acknowledge C. H. Reynolds’s assistance in utilizing these parameterizations. Available force field specific parameterizations are used by default unless the program is instructed otherwise (see arg2 below).

Calculations with solvation use periodically updated constant area and/or polarization derivatives to speed the calculation. Default update frequencies are given below. These frequencies can be changed via arg3 and arg4. If difficulties in achieving low gradients are found or if dynamics in solvent is unstable, reduce these numbers (e.g., to 2). Energy minimizations and molecular dynamics simulations using continuum solvation models 2 and 3 run approximately 1/2–1/4 the rate of in vacuo calculations.

Note: MacroModel carries out energy minimizations with an analytical, approximate function for surface areas. Thus, intermediate energies reflect the approximate function. The final energies reported, however, use an accurate numerical function. Thus, intermediate and final energies will differ.

It is recommended that the EXNB opcode be used in conjunction with solvent model 3 (GB/SA) and that the electrostatic cutoff distance be set to 20 Angstroms.

arg1

Solvation model

2

Total solvation based on approximate solvent accessible surface areas (Scheraga’s parameters).

3

GB/SA Solvation Model. Cavity and Van der Waals components from approximate solvent accessible surface areas, and electrostatic (polarization) component from GB mode. See Still et al. [21] for a discussion of effective Born radii calculation. This is the best solvent model to use.

arg2

Solvent
 

If arg2 < 0 use the general parameterization from the .slv file rather than the parametrization for the specific force field.

1, −1

Water (models 2 and 3)

5, −5

CHCl3 (model 3)

9, −9

Octanol (model 3)

arg3

Surface-area derivative update frequency

Note: This is used in models 2 and 3. Default: 25 for PRCG, SD, and OSVM minimization modes and molecular dynamics, and 1 for FMNR and TNCG. Set to 1 or 2 for problem minimizations using SD, PRCG or OSVM; the default value cannot be overridden for FMNR and TNCG.

arg4

Solvent polarization reset frequency
 

Uses constant long range components except during resets. Default: 10.

arg5

Minimum solvent/solute distance
 

Used to remove overlapping solvent molecules in model 1. (Default: 2.5 Å)

arg6

Flat-bottom positional constraint force constant
 

Used for restraining solvent molecules in model 1. (Default: 10. kcal/mol-Å2)

arg7

Half-width of flat-bottomed positional constraint
 

Used for restraining solvent molecules in model 1. (Default: 2.5 Å)

arg8

Maximum distance from solute centroid
 

Used in model 1. Molecules beyond this distance will have flat-bottomed constraints as defined in arg6 and arg7. (Default 0.0 Å)

LOGP — LOG P partition coefficient estimation

This command instructs the program to estimate the logarithm of the partition coefficient between two solvents, log Psolv1,solv2, for a series of molecules using the relationship:

where Psolv1,solv2 = [Solute]solv1 / [Solute]solv2, ΔGsolvN is the free energy of solvation of the molecule in solvent N, R is the gas constant, T is the temperature in kelvin. The first solvent, solv1, must be specified by an earlier SOLV command. The second solvent, solv2, is specified by arg2 as described below.

LOGP must be followed by a BGIN/END loop containing appropriate READ, AUTO, and MINI opcodes. The AUTO opcode should set arg6 = 1.0000 for serial calculations and arg2 = −1 to avoid generation of lists of comparison atoms. Each molecule in the input file is minimized twice, once in each solvent. When LOGP calculation is enabled the behavior of subsequent READ statements is modified and alternates between reading in new molecules and switching the solvent. Information on the free energy of solvation is collected at subsequent MINI commands. MacroModel Example: Partition Coefficient Estimation has an example of a LOGP calculation with a .com file.

The solvation models are parametrized for ambient conditions. As the temperature begins to deviate significantly from such conditions the solvation energy estimates and hence the calculated log Psolv1,solv2 become less reliable.

arg1

LOGP on/off

−1

Turn off log P estimation and generate a report.

0

Turn on log P estimation.

arg2

Solvent
 

Like arg2 of SOLV, if arg2 is less than 0 then use the default parametrization from the .slv file.

1, −1

Water (models 2 and 3).

5, −5

CHCl3 (model 3).

9, −9

Octanol (model 3).

arg5

Temperature

0.0

Use the current temperature for log P calculations (initially 298.15).

> 0.0

Use this temperature for log P calculations.

Related DEBG flags: 530 and 531.

CHGF — CHarGe File

This command causes the atomic charges used in a MacroModel energy-related calculation to come from the input structure file. If CHGF is not used, then standard charges are computed according to data in the force field file.

arg1

Source of charges

−1

Turn CHGF off.

0

Use atomic charges from input structure file (default).

1

Same as −1.

2

Use formal charges for mono-nuclear ions from the input structure file.

3

Use formal charges for all atoms and bond orders for all bonds from the input structure file.

4

A combination of 0 and 3. Use charges, formal charges for all atoms and bond orders for all bonds from the input structure file.

arg2

Treatment of sp3 CHn groups in GB solvation.
 

Different charge sets can be specified for Coulombic and for GB solvation calculations. The charges used for Coulombic calculations are written to the first charge column in the output file; the charges used for GB calculations are written to the second charge column.

Note: The only place in which this facility is currently used is in the treatment of sp3 CHn groups. When charges are assigned by the force field, then, for the purpose of GB calculations only, charges on hydrogens in such a group are added to the charge on the carbon, and the entire group is treated as a united atom.

When reading an input file, the values in the first charge column are used for Coulombic calculations and those in the second charge column are used for GB.

0

(Default.) If all H atoms of an all-atom CHn group have zero charge, unite the group for GB calculations; otherwise, treat these H atoms explicitly. If a file is written out containing force-field charges, this default recaptures the force-field behavior should the file subsequently be read in with CHGF in effect.

1

Never unite all-atom sp3 CHn groups for GB.

2

Always unite all-atom sp3 CHn groups in GB. This allows a structure with equal charge columns to be read in with CHGF and for the default behavior to be embodied in the output; that is, the output will be suitable for reading with CHGF arg2=0.