# Setting potential coefficients on a per-bead basis

Hello Everyone,

First of all, let me say that I am new to lammps. I will be conducting course grained simulations of protein fragments using a hybrid potential of charrm angles, charrm and helical dihedrals, LJ and Mercedes-Benz. One thing about my model is that the coefficients for the potentials vary based on each residues native secondary structure. (i.e. I need to specify, for each residue different coefficients).

The lammps documentation refers to specifying different coefficients for each potential type, but I am interpreting this as if the only one set of coefficients for dihedral type charrm and one for dihedral type helix not

Dihedral type charrm – alpha-helix: A = A1 d=d1

Dihedral type charrm - beta-sheet: A = A2 d=d2

Dihedral type charm – turn: A = A3 d = d3

Dihedral type helix – alpha-helix: A=A4 B=B1 C=C1

Dihedral type helix – beta-sheet: A=A5 B=B2 C=C2

Dihedral type helix – turn: A=A6 B=B3 C=C3

Some of the coefficients also depend on bead i-j pairings. Is there any way for me to specify the coefficients for each residue / residue pairing?

TIA,

Rob

Hello Everyone,

First of all, let me say that I am new to lammps. I will be conducting course grained simulations of protein fragments using a hybrid potential of charrm angles, charrm and helical dihedrals, LJ and Mercedes-Benz. One thing about my model is that the coefficients for the potentials vary based on each residues native secondary structure. (i.e. I need to specify, for each residue different coefficients).

The lammps documentation refers to specifying different coefficients for each potential type, but I am interpreting this as if the only one set of coefficients for dihedral type charrm and one for dihedral type helix not

That is not correct.

to be more explicit.

each dihedral (or bond, angle etc.) is assigned a type number and then
for each type number, you can assign a set of coefficients (and in the
case of hybrid styles also a style). you can have as many type numbers
as you want (well, up to about 2 billion IIRC).

Oh that's great, that allows me to get up and running much faster.

Thanks Axel!

Rob