[gmx-users] Improper dihedrals, which way is correct?

[Vitaly V. Chaban]

Hi,

I have got confused with setting up improper dihedrals in the gromacs
topologies. In the manual (ver. 4.0) I see the section named [
impropers ] (page 104)
"
[ impropers ]
C O1 O2 CA
"

but in the example system of the peptide
(share/gromacs/tutor/speptide) one finds defining impropers using the
second (in the same file) [ dihedrals ] section.
[ dihedrals ]
;  ai    aj    ak    al funct            c0            c1
c2            c3
    1    23     5     6     1    improper_X_X_C_X
    1    23     5     7     1    improper_N_A_C1_C
    2     5     1     3     1    improper_X_X_M3_X
    2     5     1     4     1    improper_X_X_M3_X

What is the correct way to work with improper dihedrals? I have four
sites, then angle value and then one force constant. What should my
topology line look like to accomplish that?

I would also be very grateful if somebody could share some topology
with improper angles to understand how it works.

Vitaly