AsiteDesign combines PyRosetta modules with enhanced sampling techniques to both design catalytic residues, as well as, perform in silico directed evolution of an existing active site.

A set of Python libraries are necessary to run the scripts of this Git repository; mainly OpenMM, PyRosetta, and Biotite. An easy way to install these dependencies is by using a yaml file with all the required packages to create a conda environment to use AsiteDesign. The PyRosetta credentials of the user must be added to the yaml file ("$USER" and "$PASS" must be changed). The command should be the following:

conda env create -f asitedesign.yaml 

Then, activate the conda environment:

conda activate asitedesign

Finally, you have to execute the Setup Python script to use the algorithm:

python Setup.py install

If the user wants to use the scripts for analysis and plotting (for instance, PlotSimulation.py), you must install the following dependencies in the created environment:

conda install matplotlib seaborn scikit-learn

• Python Python 3.7 or higher must be installed.
• PyRosetta (licensed software), Biotite, MPI4py, OpenMM.

To run AsiteDesign, you need the following:

• An input PDB file with the complex (ligand docked to the protein, otherwise the ligand can be placed by the code, but it's better to have it already bound)
• The parameters of all ligands, cofactors, and non-conventional amino acids that appear in the simulation (to generate them, for instance, save the ligand as mol2 file from Pymol. Then, use the facilitated molfile_to_params Python script to transform to a params file)

To start using it, the parameters that should be set in the input yaml file are the following:

• PDB --> Add the name of your input PDB file
• ParameterFiles --> Add the name of all the used params files (list them with "-")
• Name --> Add the name of the desired output folder
• DesignResidues --> Add the list of residues allowed to be mutated (ZZ leaves the residue as frozen. ZX stands for not mutable, but repackable. XX stands for mutable and repackable and XX+ adds the option to use the native residue as well. You can also specify to which residues you want to allow it to mutate by listing them, for instance, 100-A: AILFWVPY)
• CatalyticResidues --> Specify the number of residues of the active site that wants to be added (RES1, RES2 ... RESN: H)
• Ligands, 1-L (you have to specify the ligand by giving the residue number and the chain of the specific LIG). Also, the torsions that want to be excluded must be specified by the user ("ExcludedTorsions")
• Constraints --> Add the distance and sequence constraints that you want. The distance constraints should be added by passing two residues (with residue_number-chain) and two atoms (atomname) and to which values you want to constraint them (lb: value in angstroms, hb: value in angstroms)
• nIterations --> Number of adaptive sampling epochs that want to be performed
• nSteps --> Number of steps performed in each epoch/iteration
• nPoses --> Number of final poses (mutants/designs) to be reported (each one given to a processor/CPU)
• Time --> Time in the queue (if it's run in a cluster)

To get a better understanding, check the examples (both yaml files) in the Templatized_control_file folder

If AsiteDesign is used in one of your publications, please cite the paper: https://doi.org/10.1021/acs.jpcb.2c07091

For any queries, feel free to ask us:

Masoud Kazemi, (@masoudk), Biomatter Designs, Vilnius, Lithuania Sergi Roda, (@SergiR1996), Barcelona Supercomputing Center, Barcelona, Spain