Ucsfdock

ucsfdock is a Python package wrapping the DOCK program that provides tools to help standardize and automate the computational methods employed in molecular docking.

Programs:

• blastermaster: generate a specific docking configuration for a given receptor and ligand
• dockmaster: evaluate many different docking configurations in parallel using a specified job scheduler (e.g. Slurm)

A docking configuration is a unique set of DOCK parameter files and INDOCK parameter values.

Installation

1. TODO

Instructions

blastermaster

blastermaster allows the generation of a specific docking configuration for a given receptor and ligand.

Invoking blastermaster commands below will produce a log file called blastermaster.log in your current working directory.

blastermaster configure

First you need to create the directory for your blastermaster job. To do so, simply type

blastermaster configure

By default, the job directory is named blastermaster_job. To specify a different name, type

blastermaster configure <JOB_DIR_NAME>

The job directory contains two sub-directories:

1. working: input files, intermediate blaster files, sub-directories for individual blastermaster subroutines
2. dockfiles: output files (DOCK parameter files & INDOCK)

If your current working directory contains any of the following files, then they will be automatically copied into the working directory within the created job directory.

• rec.pdb
• xtal-lig.pdb
• rec.crg.pdb
• reduce_wwPDB_het_dict.txt
• filt.params
• radii
• amb.crg.oxt
• vdw.siz
• delphi.def
• vdw.parms.amb.mindock
• prot.table.ambcrg.ambH

This feature is intended to simplify the process of configuring the blastermaster job. If you would like to use files not present in your current working directory, copy them into your job's working directory, e.g.:

cp <FILE_PATH> <JOB_DIR_NAME>/working/

Finally, configure the blastermaster_config.yaml file in the job directory to your specifications. The parameters in this file govern the behavior of blastermaster.

blastermaster run

Once your job has been configured to your liking, navigate to the the job directory and run blastermaster:

cd <JOB_DIR_NAME>
blastermaster run

This will execute the many blastermaster subroutines in sequence. The state of the program will be printed to standard output as it runs.

dockmaster

dockmaster allows the evaluation of many different docking configurations in parallel using a specified job scheduler (e.g. Slurm).

The name "dockmaster", aside from being an uncreative rehash of the name "blastermaster", derives from the notion of a literal dockmaster, i.e., the person in charge of a dock who manages freight logistics and bosses around numerous dockworkers. In this analogy, a single dockworker corresponds to the processing of a single docking configuration.

Invoking dockmaster commands will produce a log file called dockmaster.log in your current working directory.

dockmaster configure

First you need to create the directory for your blastermaster job. To do so, simply type

dockmaster configure

By default, the job directory is named dockmaster_job. To specify a different name, type

dockmaster configure <JOB_DIR_NAME>

The job directory contains two sub-directories:

1. working: input files, intermediate blaster files, sub-directories for individual blastermaster subroutines
2. retro_docking: individual retro docking jobs for each docking configuration

The key difference between the working directories of blastermaster and dockmaster is that the working directory of dockmaster may contain multiple variants of the blaster files (prefixed by a number, e.g. "1_box"). These variant files are used to create the different docking configurations specified by the multi-valued entries of dockmaster_config.yaml. They are created efficiently, such that the same variant used in multiple docking configurations is not created more than once.

If your current working directory contains any of the following files, then they will be automatically copied into the working directory within the created job directory.

• rec.pdb
• xtal-lig.pdb
• rec.crg.pdb
• reduce_wwPDB_het_dict.txt
• filt.params
• radii
• amb.crg.oxt
• vdw.siz
• delphi.def
• vdw.parms.amb.mindock
• prot.table.ambcrg.ambH

This feature is intended to simplify the process of configuring the dockmaster job. If you would like to use files not present in your current working directory, copy them into your job's working directory, e.g.:

cp <FILE_PATH> <JOB_DIR_NAME>/working/

Finally, configure the dockmaster_config.yaml file in the job directory to your specifications. The parameters in this file govern the behavior of dockmaster.

environmental variables

Designate where the short cache and long cache should be located.

export SHRTCACHE=/dev/shm
export LONGCACHE=/dev/shm

In order for dockmaster to know which scheduler it should use, please configure the following environmental variables according to the options available.

Slurm

E.g.:

export QSTAT_EXEC=/usr/bin/sbatch
export SQUEUE_EXEC=/usr/bin/squeue

SGE

E.g.:

export QSTAT_EXEC=/opt/sge/bin/lx-amd64/qstat
export QSUB_EXEC=/opt/sge/bin/lx-amd64/qsub

The following is optional:

export SGE_SETTINGS=/opt/sge/default/common/settings.sh 

dockmaster run

Once your job has been configured to your liking, navigate to the the job directory and run dockmaster:

cd <JOB_DIR_NAME>
dockmaster run <JOB_SCHEDULER_NAME>

where <JOB_SCHEDULER_NAME> is one of:

• sge
• slurm

This will execute the many dockmaster subroutines in sequence, including the submission of retro docking jobs for all docking configurations. The state of the program will be printed to standard output as it runs.

Once the dockmaster job is complete, the following files will be generated in the job directory:

• dockmaster_job_report.pdf: contains (1) roc.png of best retro docking job, (2) box plots for every multi-valued config parameter, and (3) heatmaps for every pair of multi-valued config parameters
• dockmaster_job_results.csv: enrichment metrics for each docking configuration

In addition, the best retro docking job will be copied to its own sub-directory best_retro_docking_job/.

Within each retro docking job directory, there are the following files and sub-directories:

• working/: intermediate files
• dockfiles/: parameters files and INDOCK for given docking configuration
• output/: contains sub-directories 1/ for actives and 2/ for decoys, each containing OUTDOCK and test.mol2 files
• retro_docking_job_results.csv: data loaded from OUTDOCK files for both actives and decoys
• roc.png: the ROC enrichment curve (log-scaled x-axis) for given docking configuration