[gmx-users] problem with energy minimization after g_cluster

[DimitryASuplatov]

Hello,
I have created a consensus structure from the last ns of my trajectory
using g_cluster with default parameters and -av for writing average
structure.
When I am trying to energy minimize output structure with steep I get a
non-sense result - 

   Tolerance (Fmax)   =  2.00000e+02
   Number of steps    =         3000
Step=   17, Dmax= 1.5e-06 nm, Epot=  1.51233e+18 Fmax=         inf,
atom= 11754
Stepsize too small, or no change in energy.
Converged to machine precision,
but not to the requested precision Fmax < 200

This problem does not occur when using g_cluster without -av averaging -
minimization performs as usual with a conversion in about 1000 steps.

Inspecting and comparing averaged and non-averaged structures I`ve found
that 760 Leu (the one that is reported to have the highest energy) is
corrupt - averaging creased the sidechain atoms to about two or three
positions (I mean that several atoms overlap).
Is that some kind of bug?
Thanks.