[gmx-users] constraining two diff molecules removes LJ interaction?

[Mark Abraham]

>
> This is my first post to this list, so a warm hello to everybody!
>
> I want to constrain the distance between two atoms in different
> molecules (two diff proteins, in complex). So, I defined both molecules
> in the same [ moleculetype ] and added an extra [ bonds ] line:
> 216   6465     6 0.400 300.00 ;
> (notice it is type 6, for harmonic constrain). The constrain seems to
> work fine, but it looks that LJ interaction between the centers is also
> removed (or scaled) since grompp gives me LJ14:   31800 with the
> constrain and  LJ14:   31799 without it (I am guessing that this LJ14 is
> the number of 1-4 exclusions from the LJ pair list , correct?). But
> since these two centres (216 and 6465) are actually in different
> molecules, their LJ interaction should be included.
>
> The manual says ``There is a harmonic interaction ([ bonds ] type 6)
> which does not connect the atoms by a chemical bond''. So, why do I see
> an increase in the LJ14 number? Should I re-include this LJ interaction
> using [ pairs ]? Which type? Need to re-assing the C6 and C12 coeff.?

Using gmxcheck and/or gmxdump on your two .tpr files will allow you to
confirm where the problem is. Be aware that the indices that you will see
in the output from gmxdump start from zero, C-style, and those in your
coordinate file and topology file start from one.

Mark