Quick Start

• on this page, we will introduce a quick guide on running dissociation PaCS-MD using gromacs.
• If you want to try other PaCS-MD, please refer to Analyzer.

Content

• Step1: Installing by pip or conda
• Step2: Preparing initial files
• Step3: Preparing input file for PaCS-MD
• Step4: Running PaCS-MD
  • Supplement: Runing other trial
• Step5: Making new fitted trajectories from pacsmd results
• Step6: Extracting collective-variables from fitted trajectories
• Step7: Building MSM and predicting free energy

Step1: Installing by pip or conda

1. Running git clone for getting source code

git clone https://github.com/Kitaolab/PaCS-Toolkit.git

2. Installing with your favorite method

# Install by pip
pip install -e ".[mdtraj]"

# Or install by conda and pip
conda create -n pacsmd "python>=3.7" -y
conda activate pacsmd
pip install -e ".[mdtraj]"

• The above methods are better, but sometimes they are not suitable for a special situation.
• For more information about installation, Please refer to Install page.

Step2: Preparing initial files

• On this page, we will use gromacs as simulator.
• In gromacs, gro, top, mdp and ndx files are necessary. Please get these files ready.
• Sample jobscripts and input files are here

Step3: Preparing input file for PaCS-MD

• PaCS-MD requires a toml format input file. Please make input.toml
  • Please adjust the above settings as needed.
  • For more information about this file, please refer to input file.

vim input.toml

# Input file for PaCS-MD
## basic
# pacsmd settings
## basic
max_cycle = 2                           # Maximum number of cycles to run. (ex. 1, ..., 123, ..., 999)
n_replica = 3                           # Number of replica. (ex. 1, ..., 123, ..., 999)
n_parallel = 3                          # Number of replica which are calculated at a time
skip_frame = 1                          # Frequency of frames used for ranking among trajectories
centering = true                        # Whether to move the molecule to the center
centering_selection = "protein"         # Name of molecule to move in the center
working_dir = "./."                     # Directory where pacsmd will run

## simulator
simulator = "gromacs"                   # Software used inside PaCS-MD
cmd_mpi = "mpirun -np 4"                # Commands for MPI such as mpirun, blank is OK
cmd_serial = "gmx_mpi mdrun -ntomp 6"   # Commands to run the simulator serially
cmd_parllel = "gmx_mpi mdrun -ntomp 6"  # Commands to run the simulator parallelly
structure = "input.gro"                 # Structural file such as gro, pdb, rst7, etc.
topology = "topol.top"                  # Topology file such as top, parm7, psf, etc.
mdconf = "parameter.mdp"                # Parameter file such as mdp, mdin, namd, etc.
index_file = ".index.ndx"               # Gromacs index file
trajectory_extension = ".xtc"           # Trajectory file extension. ("." is necessary)

## analyzer
type = "dissociation"                   # Evaluation type
threshold = 100                         # CV threshold for determining to terminate a trial
skip_frame = 1                          #  How many frames to skip when ranking CVs
analyzer = "mdtraj"                     # Trajectory tool used to calculate the evaluation type
selection1 = "resid 1 to 5"             # Selection string for specified group in trajectories
selection2 = "resid 6 to 10"            # Selection string for specified group in trajectories

## postprocess
genrepresent = true                     #  Whether genrepresent is executed after trial
rmmol = true                            #  Whether rmmol is executed after each cycle
keep_selection = "not water"            #  Molecular name or index group to be left in the trajectory when rmmol
rmfile = true                           #  Whether rmfile is executed after trial

Step4: Running PaCS-MD

• Finally, we are ready to perform PaCS-MD
• Specifying the trial id in argument t and the input file in argument f.

pacs mdrun -t 1 -f input.toml

• If you want to continue the simulation from the middle of a cycle or trial, simply run pacsmd again. Completed cycles will be skipped, and the simulation will resume from the the point of interruption.

• You will get the results

$ ls
trial001/

Supplement: Runing other trial

• Usually, a single trial is not sufficient for sampling, so multiple trials are required.
• We can run trial2 PaCS-MD

pacs mdrun -t 2 -f input.toml

• Output will be the following

$ ls
trial002/

Step5: Making new fitted trajectories from pacsmd results

• For some analyses, fitting trajectories can make subsequent calculations smoother.
• For fitting, we prepare pacs fit mdtraj command. But you can use existing software such as gmx trjconv or cpptraj as usual.

$ pacs fit trial mdtraj -t 1 -tf prd_rmmol.xtc -top rmmol_top.pdb -r ref.gro -ts protein -rs protein

Step6: Extracting collective-variables from fitted trajectories

• After making fitted trajectories, you need to extract CV to build MSM, where CV denotes collective variables such as "distance" and "inter COM vector" and "PCA" and etc.
• For extracting CVs, we got pacs genfeature command. But this command provides you only frequently used CV such as com-distance, com-vector and rmsd.
• So if you want to use other specific CVs, you need to write a code by yourself.

$ pacs genfeature comdist mdtraj -t 1 -tf prd.xtc -top inputs/example_gromacs/input.gro -s1 "residue 1" -s2 "residue 9" 
$ ls
comdist-CV/

Step7: Building MSM and predicting free energy

• After extracting CVs, various analyses can be performed on them.
• PaCS-MD is especially compatible with analyses using MSM.