[AMBER] Umbrella Sampling and WHAM analysis

From: mish <smncbr.gmail.com>
Date: Mon, 18 Feb 2019 11:29:31 +0000

Hello everybody,

I've a question regarding umbrella sampling (US) with distances and further
analysis with the WHAM program of Alan Grossfield. I ran 1D US simulations
of two a Tryptophan-small-molecule interaction, where I am pulling them
away in explicit solvent condition. Input file for a window at 4A distance
looks like:
3ns NPT production for angstrom
  imin = 0, ntx = 5, irest = 1,
  ntpr = 10000, ntwr = 20000, ntwx = 20000,
  ntf = 2, ntc = 2, cut = 10.0,
  ntb = 2, nstlim = 1500000, dt = 0.002,
  temp0 = 300, ntt = 3,
  gamma_ln = 1.0,
  ntp = 1, pres0 = 1.0, taup = 5.0,
  nmropt = 1, ioutfm=0, iwrap=0
  type='DUMPFREQ', istep1=1000,

   r1=1, r2=04.0, r3=04.0, r4=17,
   rk2=1.0, rk3=1.0,

Then I am using WHAM to get PMF
wham 1 17 50 0.00000001 300 0 wham.in wham.out 10 9999
where wham.in file has entries for each 13 windows:
win-03.0/dist01_03.0.dat 3.00 2.00
win-15.0/dist01_15.0.dat 15.00 2.00

Histograms looks perfectly fine, but the PMF shows me a strange decrease
after 8A while being pulled away further. I used a TIP4PEWBOX extending 30A
from the solute to make sure that solute going of of the box is not causing
any problem. I have tried many more replicas and longer simulations to
ensure sampling but results are same as here.

I have no clue why PMF starts decreasing while two molecules are separated
by 8A and there is nothing else to interact other than water molecules.
There are several -OH in small-molecule but I believe they should have a
minima while interacting to each other. I am including histogram (upper
panel) and PMF (lower) from the simulation.

Can someone point-out the possible cause of such a behavior?
[image: image.png]

AMBER mailing list

(image/png attachment: image.png)

Received on Mon Feb 18 2019 - 04:00:02 PST
Custom Search