SECTION 1

Umbrella equilibration of initial structure

Before starting a TASS simulation, we need to carry out equilibration run at 300K for every umbrella window for about 100 ps. The purpose of this run is to break the reaction coordinate (in our case dihedral angle Φ1) up into the series of windows and then applying the restraint that acts to force the reaction coordinate to remain close to centre of window. The selection criteria for number of umbrella windows and the force constant of umbrella restraint is exactly similar to that of umbrella sampling i.e. the number of windows must be such that every window must sample some space of next window. Similarly, force constant must be large enough to sample only the small subset of space we want to sample but it should not be too large that there is no overlap between the two consecutive windows.

We will apply umbrella bias along Φ1 from –π to π (-3.2 rad to 3.2 rad for simplicity) at an interval of 0.2 rad. Till now, we have one starting structure (say Φ1= -1.4 rad). If we apply restraint on it, to sample whole space of Φ1 then, in some part of space there will be sudden change in conformation due to sudden change in forces. To avoid this situation, we will start simulation with structure at Φ1= -1.4 rad and use the restart (.rst file) from this as the input of -1.6 rad simulation and so on, upto -3.2 rad simulation. Since the value of dihedral angles are periodic, - 3.2 and 3.2 rad have same conformation. Therefore, we will use restart file of 3.2 rad simulation as input of 3.0 rad simulation and so on, upto the -1.2 rad simulation. In this way we will sample Φ1 by placing 33 restraining umbrella potentials.

The following input files are required for each umbrella equilibration:

  1. ala_prmtop : It describes the parameter and topology of the molecules of the system
  2. ala_prod.rst : It describes the coordinates of the molecules of the system
  3. mds.in : It describes the input settings for Amber MD engine
  4. plumed.dat : It describes the input settings for PLUMED-AMBER interface

mds.in

Alanine Tri Peptide
 &cntrl
  irest=0, ntx=1,
  imin = 0, ntb = 0,
  igb = 0, ntpr = 1000, ntwx = 1000,
  ntt = 3, gamma_ln = 1.0,
  tempi = 300.0, temp0 = 300.0,
  nstlim = 100000, dt = 0.001,
  nscm=100,
  cut=999.0, rgbmax=999.0,
plumed=1,
plumedfile='plumed.dat'
&end

plumed.dat

#set up variable for Phi dihedral angle 
phi1: TORSION ATOMS=5,7,9,15
#psi1: TORSION ATOMS=7,9,15,17
#phi2: TORSION ATOMS=15,17,19,25
#psi2: TORSION ATOMS=17,19,25,27
####################################################
# Impose an umbrella potential on CV1 

restraint-phi1: RESTRAINT ARG=phi1 KAPPA=500 AT=XXXX
####################################################
# monitor the variable phi
PRINT STRIDE=10 ARG=phi1 FILE=COLVAR

Here, we need to replace XXXX by centre of umbrella window we want to sample. Now, we can run each simulation through sander. For example to run them using 2 processors :

mpirun -np 2 $/AMBERHOME/bin/sander.MPI -O -i mds.in -o umb.out -p ala_prmtop -c ala_prod.rst -r umb.rst -x umb.mdcrd

The output files of these simulations are:

  1. umb.out : output file
  2. umb.rst : output restart file with coordinates
  3. umb.mdcrd : Output trajectory file for MD simulation
  4. COLVAR : Collective Variable file

This will generate 33 structures relaxed at the centre of every umbrella window which will become input structures of TASS simulation.

Go through the amber tutorial and plumed tutorial for basic understanding of keywords used in mds.in and plumed.dat files.