rizzolab / dock6_screening_protocols Goto Github PK

This is the collection of scripts and parameters that make up the Virtual Screening Protocol for the Rizzo Lab at Stony Brook University.

NOTE BEFORE PROCEEDING: 
These scripts have been tested in-house, but are often not the complete picture for each system's experiment or needs. Modification from this framework may often be necessary for a given system. Use at own risk.

If any bugs are encountered, please contact the Rizzo Lab.


Programs required to run these scripts:
    DOCK6 and accessories
    AMBER (for tleap, antechamber, sander)
    AMBERTools
    MOE

Not every step requires every piece of software.

Directories
    zzz.master
        - Where all of the files prepared before beginning should go.
    zzz.parameters
        - Collection of parameters that are used by the main scripts.
    zzz.rank_scripts
        - Collection of scripts used to rank outputs by various scores. Referred to as
          run.008 etc. Denovo and virtual screening scripts denoted separately.
    zzz.scripts
        - Collection of ancillary scripts called at various stages.


The Main Scripts:
run.000.set_env_vars.csh
    - Sets the environment variables to be used in the rest of the scripts.
    - Must be edited to match your own environment.
    - Must be sourced before other scripts will function.

run.001.lig_clean_am1bcc.csh
    - Takes the prepared cognate ligand and performs the following:
        - Pre-processes for use with DOCK.
        - Charges it with AM1-BCC charges using Antechamber, from the Amber directory.
        - Also does the same with any prepared cofactor.

run.002.rec_runleap.csh
    - Prepares the receptor for generating a grid and for use in MD.
    - Removes any hydrogens present and reprotonates and charges with tleap (Ambertools)
    - Checks for any long bonds, disulfide bonds, and HIE/HID residue naming.
    - Generates and minimizes a complex with the receptor, ligand, and cofactors.
    - Separates the minimized ligand from the complex.

run.003.rec_dms_sph.csh
    - Runs dms on the receptor to generate a DMS surface to generate spheres.
    - Runs sphgen on the dms surface outfile to generate spheres for use in docking.

run.004.rec_grid_cluster.csh
    - Generates a box and runs grid to generate an energy grid that will be used in docking.\

run.005.vendor_make_subsets
    - Breaks a library of molecules into smaller chunks suitable for virtual screening on a cluster.
    - Also does some pruning for min/max rotable bonds and absolute net charge (user defined).

run.006a.dock_to_grid.csh
    - Writes docking files for each chunk.
    - Writes a submission script for SLURM.
    - Submits the SLURM script w/ user specified timing and node information.

run.006b.dock_leftovers.csh
    - Determines which molecules did not dock (wall clock limit, job died unexpectedly, etc).
    - Concatenates these molecules to a file and submits a new docking job for them.

run.007a.cartesian_min.csh
    - Takes the output from docking to a grid and docks and rescores each molecule in Cartesian space.

run.007b.make-mg.sh
    - Takes the primary residues found in 007a and generates a multigrid based on that output.

run.008.submit_rankjobs.csh
    - Submits jobs to a cluster to rank the final output by each score.

run.009.descriptor_rescore.ranked.csh
    - Rescores each ligand based on a footprint reference.
    
run.010.moe_postprocess.csh
    - Postprocessing script for virtual screening. 
    - Calculates descriptors and clusters molecules by footprint.

run.011.write_mol2_fp_updated.csh
    - Writes output ranked files of clusterheads ranked by scoring function, which should be visually inspected for compound selection.
    - Generates footprint plots for each molecule.


That concludes the traditional virtual screening protocol. The rest of the scripts deal with an expanded de novo protocol for building brand new molecules in a fragment-based fashion. 


run.012a.dn.submit_denovo.csh
    - Generates input files for de novo docking and submits them to a cluster
    - Requires specifying the anchor directory, which is a user defined collection of anchors to use as a base in de novo design.
    
run.012b.dn.prune_prepare_cartmin.csh
    - Extracts a top scoring copy of each molecule from multiomol2 output of the previous step.
    - Submits a cartesian minimization to the cluster.

run.013.dn.submit_rankjobs.csh
    - Same function as run.008, for de novo molecules.

run.014.dn.descriptor_rescore.ranked.csh
    - Same function as run.009, for de novo molecules.

run.015.dn.moe_postprocess.csh
    - Same function as run.010, for de novo molecules.

run.016.dn.ping_databases.csh
    - Similar function to run.011, for de novo molecules.
    - Will also ping ZINC and PubChem for a similarity search (hardcoded cutoff being 0.75) for the molecules to see if there are analogues available.