In this page, we will run a SMD simulation on 1E0Q.
PATH: ~/1E0Q/amber/SMD/
copy 1E0Q.pdb, min.mdin, heat.mdin, md.mdin to ~/1E0Q/amber/SMD/
./pipeline.sh
the only two files we will be using are 1E0Q_md.rst and 1E0Q_box.prmtop
let's run 100A/ns SMD simulation as an example:
Input File Preparation
mkdir 100_ang_per_ns
copy create_ASMDinputs.sh and create_job.sh to ./100_ang_per_ns/
create_ASMDinputs.sh:
#!/bin/bash args=("$@")
#check the correct num of args
num_asmd_sim=${args[0]}
#create the ASMD directories
#create the distance RST files
for stage in {1..1}; do
&end |
./create_ASMDinputs.sh 5 250000
creat_job.sh
#!/bin/bash args=("$@")
#check the correct num of args num_asmd_sim=${args[0]}
stage=${args[2]}
cat >>_job.sh<<EOF
do_parallel="pmemd.cuda" EOF
for ((counter=1;counter<=$num_asmd_sim;counter+=1)); do sed 's/SSS/$/g' _job.sh > job.$stage.sh; rm -f _job.sh |
./creat_job.sh 5 ../1E0Q_md.rst 1
Run the simulation
./job.1.sh
Repeat the same procedure for 10A/ns.
./create_ASMDinputs.sh 5 2500000
./creat_job.sh 5 ../1E0Q_md.rst 1
Data Analysis
./10ang_per_ns/:
../ASMD.py -i asmd*.work.dat.1 -o jar.stage1.dat
./100ang_per_ns/:
../ASMD.py -i asmd*.work.dat.1 -o jar.stage1.dat
plotting the PMF:
gnuplot:
set title "PMF of Steered Molecular Simulation of 1E0Q"
set xlabel "end-to-end distance(A)"
set ylabel "PMF (kcal/mol)"
plot "10ang_per_ns/jar.stage1.dat" w l ti "100A/ns","100ang_per_ns/jar.stage1.dat" w l ti "10A/ns"
As we can see, PMF for faster speed is significantly bigger than those of slower pulling speed, That is the reason when we were trying to explore the difference of gibbs free energy between two confomer using PMF, we need to emply as slow pulling speed as possible so as to make it close to reversible work.
Reference: