[lammps-users] consistency of units in MD simulations with hybrid pair_style

Hi,

I would like to run a simulation of a peptide bound to a gold surface. I
intend to use a hybrid force field comprised of an AMBER force field for
the peptide, an EAM potential for the gold atoms and a LJ potential for
the interaction of the gold atoms with the peptide. Thus I define the pair
style as

pair_style hybrid lj/charmm/coul/long 8.0 10.0 eam lj/cut 5.0

I have written an input file that works (no parse errors). I have also
employed Antechamber and xLeap to obtain the AMBER parameters for the
peptide. If I am not wrong, the AMBER parameters are given in "real"
units. Therefore, I have set "units real" in my input file. However, I
also use the file "Au_u3.eam", which contains information in "metal"
units. Thus I would like to know what I should do in such a way that
LAMMPS gets all the force field parameters in the correct units either I
set units = "real" or "metal". That is,

1) Should I convert the parameters in Au_u3.eam to "real" units and set
"units real" in the input file? Is that enough? Is there any tool to do
that?

2) Instead, should I convert the AMBER parameters to "metal" units and set
"units metal" in the input file? Is it enough?

3) Is there any other simpler way to run this kind of simulations with
both AMBER parameters (in real units) and EAM parameters (in metal
units)?.

I would appreciate advice about which is the best option to run this kind
of simulations.

regards,

Carlos