[gmx-users] Umbrella sampling - abnormal potential energy profile and wrong PMF profile

[gozde ergin]

Dear gromacs users,

My simulation system has 512 water molecules covered by 25 organic
molecules on each side (1st pic). I have also gas phase.

Link for the plots :
http://imgur.com/HobCa8H,shH3pA9,OZpCUTZ,DeDk35Q,iFDGzXW,67SHwUw,Jj572Wx,2E1TJuH,xNC3D1s,JChHqjO,Z5Uu1l1,2mKA8UO,9DSe8OO#0

In order to capture the PMF profile of the system on z direction, I
selected a one water molecule from the middle of the bulk and pulled it to
gas phase.

For very long time I am trying to understand why do I have this abnormal
potential energy profile during umbrella sampling simulation (10th pic) and
pressure profile (11th pic).

This potential energy is for the window when pulled molecule inside the
bulk.

Also when I  extract the PMF from 18 to 20 ns which is flat region (g_wham
-it itp.dat -if pullf.dat -o -b 18000 -e 20000)  this PMF does not seem
correct (12 and 13th pics)

I did the same simulation by using different organic but I get the same
result! (abnormal potential, pressure and PMF profile)

I am doing NVT simulation after minimization and I *do not do NPT*
simulation before the NVT.

Maybe I should?
Does anyone face with the same problem?
Do you have any advice?

Thanks in advance.