bieniekmateusz / somd-rbfe-protocol

Up to date protocol for preparing and running RBFE simulations (Cole's group)

Relative binding free energy - tutorial

RBFE input files were taken from the sarscov2-3toR study case in the fegrow package.

BioSimSpace protocols and scripts used to generate the input files for free energy MBAR calculations.

The following details the procedure for protein-ligand binding free energy calculations.

File Setup for Free Energy Calculations

Install BioSimSpace (https://github.com/michellab/BioSimSpace/)

Step 1a) Parameterise the protein with AMBER forcefield:

• Given the protein in pdb format, use the parameterise.py script from BioSimSpace to parameterise it with the ff14SB amber forcefield, and generating .rst7 and .prm7 files: parameterise.py --input PROT.pdb --forcefield ff14SB --output PROT
• save the generated .rst7 and .prm files in the 1.prot_param directory

Step 1b) Generate AMBER format forcefield for the ligands (GAFF):

• Given a molecule in a pdb format in 0.lig_input directory, use parameterise.py --input MOL.pdb --forcefield GAFF2 --output MOL
• Save each molecule's topology .prm and coordinates .rst7 in 1.lig_param

Step 1c) Combine each ligand and protein:

• for each ligand, combine it with the protein: combine.py --system1 MOLX.prm7 MOLX.rst7 --system2 PROTEIN.prm7 PROTEIN.rst7 --output PROT_MOLX in the 1.com_param directory
• this will create unsolvated .prm7 and .rst7 files

Step 2) Generate the files for the free energy calculations:

• create the perturbation files for the free energy calculations. E.g. define the transition of Mol1 to Mol2 which creates .mapping, .mergeat0.pdb. .pert, .prm7 and .rst7 files. For the unbound alchemical leg run prepareFEP.py --input1 MOL1.prm7 MOL1.rst7 --input2 MOL2.prm7 MOL2.rst7 --output MOL1_MOL2 to generate files in the forward/2.lig_fep directory. For the bound alchemical leg: prepareFEP.py --input1 PROT_MOL1.prm7 PROT_MOL1.rst7 --input2 PROT_MOL2.prm7 PROT_MOL2.rst7 --output PROT_MOL1_MOL2 and output the files in the 2.com_fep directory.

Step 3a) Solvate both unbound and bound legs:

• solvate each ligand transformation with: solvate.py --input MOL1_MOL2.prm7 MOL1_MOL2.rst7 --output MOL1_MOL2_sol --water tip3p --box_dim 35 in the 3.lig_soleq
• solvate each complex transformation with: solvate.py --input PROT_MOL1_MOL2.prm7 PROT_MOL1_MOL2.rst7 --output PROT_MOL1_MOL2_sol --water tip3p --box_dim 90 in the 3.com_soleq directory
• --box_dim is the box size used for our system in Angstroms.

3b) Equilibrate the solvated systems:

• equilibrate the bound leg (prot_mol_sol.rst7): amberequilibration.py --input MOL1_MOL2_sol.prm7 MOL1_MOL2_sol.rst7 --output MOL1_MOL2_soleq in 3.lig_soleq directory.
• equilibrate the unbound leg (mol_sol.rst7): amberequilibration.py --input MOL1_MOL2_sol.prm7 MOL1_MOL2_sol.rst7 --output MOL1_MOL2_soleq in 3.com_soleq directory.

4) Free Energy Calculations

Copy the directory Parameters and the scripts complex_lambdarun-comb.sh and ligand_lambdarun-comb.sh into the directory forward. The Parameters folder should contain the main configuration file lambda.cfg which the user should verify and modify accordingly.

For each transformation:

1. Initiate the transformation MOL1_MOL2 by running init.py MOL1 MOL2 from within the forward directory.
2. The lambda.cfg file contains various parameters, namely the number of moves and cycles, the timestep, the type of constraints, the lambda windows used and the platform on which to run the calculation.
3. Run the ligand_run.sh and complex_run.sh scripts to carry out the simulations.
4. Gather the results by runing analyse_freenrg mbar -i lambda-*/simfile.dat -o out.dat -p 90 in all discharge and vanish directories. Further calculate the corrections:

Corrections:

• FUNC.py: Evaluates the electrostatic correction for the free energy change: FUNC_corr. This is run for lambda= 0 at the discharge leg.
• lj-tailcorrection -C sim.cfg -l <lambda> -b 1.00 -r traj000000001.dcd -s 20 Evaluates the end-point correction for the truncated vdW potentials. This is run for lambda= 0 and lambda= 1 of the vanish leg.

Results

In addition to the above instructions for generating the input files for this paper, the results directory also contains the raw data (free energy calculations) which can be analysed with python run_networkanalysis.py sars/sarscov2-3toR.csv --target_compound 14 -o sars.dat -e sars/sarscov2_ic50_exp.csv --stats --generate_notebook.

This protocol might require more repetition for good sampling. Repeat the steps 3-6 with MOL2_MOL1 instead of MOL1_MOL2 in order to obtain the backward directories.