Autodock & Autodock-GPU on Biowulf

Quick Links

AutoDock is a suite of automated docking tools. It is designed to predict how small molecules, such as substrates or drug candidates, bind to a receptor of known 3D structure.

Autodock is still available on Biowulf, but we enocourage users to try the newer version, Autodock-GPU

Autodock is apart of the Autodock Suite, along with: AutodockCrankPrep and Autodock Vina

Documentation

Autodock website at Scripps

Important Notes

Module Name: Autodock version 4.2.6
Module Names: Autodock-GPU
AutodockCrankPrep is also loaded with Autodock-GPU (seeAutodockCrankPrep )
If you wish to run the mgltools Python scripts you may need to use the pythonsh command available when you load the module. This will ensure your environment is properly set up.
mgltools is a graphics-intensive package, and can easily be installed/run on your desktop system for better performance. It can be downloaded here.
Autogrid4 is now available with the Autodock-GPU module
The entire PDB (updated once a week) database is available in /pdb on all nodes, the Biowulf login node, and Helix.

Autodock - GPU

Autodock-GPU processes ligand-receptor poses in parallel over multiple compute units on GPUs. [Autodock-GPU website]

Example files for Autogrid: /usr/local/apps/Autodock/gpu/AutoGrid/Tests

Example files for Autodock-GPU: /usr/local/apps/Autodock/gpu/1.6/input

Autogrid4 and Autodock-GPU: Sinteractive session (on GPU)

[user@biowulf]$ sinteractive --gres=gpu:a100:1 --ntasks=8 --ntasks-per-core=1
[user@cn3144 ~]$ module load Autodock-GPU
[+] Loading CUDA Toolkit  12.1.0  ...
[+] Loading Autodock  1.6
[+] Loading AutodockCrankprep  1.0 on cn3144
[user@cn3144 ~]$ cp -a /usr/local/apps/Autodock/gpu/AutoGrid/Tests .
[user@cn3144 ~]$ cp -a /usr/local/apps/Autodock/gpu/1.6/input .
[user@cn3144 ~]$ autogrid4
usage: AutoGrid         -p parameter_filename
                        -l log_filename
                        -d (increment debug level)
                        -h (display this message)
                        --version (print version information, copyright, and license)
[user@cn3144 ~]$ autodock_gpu -h
AutoDock-GPU version: v1.6-release
Command line options:
Arguments              | Description                                           | Default value
-----------------------+-------------------------------------------------------+------------------
INPUT
--lfile             -L | Ligand pdbqt file                                     | no default
--ffile             -M | Grid map files descriptor fld file                    | no default
--flexres           -F | Flexible residue pdbqt file                           | no default
--filelist          -B | Batch file                                            | no default
--import_dpf        -I | Import AD4-type dpf input file (only partial support) | no default
--xraylfile         -R | reference ligand file for RMSD analysis               | ligand file
...

Autogrid4 a batch job:

#!/bin/bash
module load Autodock-GPU
cd Tests/
autogrid4 -p hsg1_sm.gpf -l hsg1_sm-output.glg

Select a GPU type (e.g. a100, v100x, v100) and submit with:

sbatch --partition=gpu --gres=gpu:a100:1  autogrid4.sh

Autodock-GPU

As a batch job:

#!/bin/bash
module load Autodock-GPU
cd /input/1ac8/derived
autodock_gpu -M 1ac8_protein.maps.fld -L 1ac8_ligand.pdbqt

Select a GPU type (e.g. a100, v100x, v100) and submit with:

sbatch --partition=gpu --gres=gpu:a100:1  autodock-gpu.sh

Autodock

We encourage users to use HPC OnDemand for launching graphical sessions

Launch a graphical session to use ADT and PMV:

[user@cn3144 ~]$ module load Autodock
[+] Loading Autodock 4.2.6  ...
[user@cn3144 ~]$ adt
setting PYTHONHOME environment
Run ADT from  /usr/local/apps/Autodock/mgltools_x86_64Linux2_1.5.7/MGLToolsPckgs/AutoDockTools
MSMSLIB 1.4.4 started on cn2063
Copyright M.F. Sanner (March 2000)
Compilation flags

At this point you should see two Autodock windows appear.

Follow the same process for PMV:

[user@cn3144 ~]$ module load Autodock
[+] Loading Autodock 4.2.6  ...
[user@cn3144 ~]$ pmv
setting PYTHONHOME environment
Run PMV from  /usr/local/apps/Autodock/mgltools_x86_64Linux2_1.5.7/MGLToolsPckgs/Pmv
MSMSLIB 1.4.4 started on cn2063
Copyright M.F. Sanner (March 2000)
Compilation flags

At this point you should the Python Molecule Viewer window appear.

Batch job

Create a batch input file (e.g. autodock.sh). For example:

#!/bin/bash
module load Autodock
autodock4 -p myfile.dpf -l myfile.dlg

Submit this job using the Slurm sbatch command.

sbatch [--mem=#] autodock.sh

The --mem=#g flag should be used if the autodock run requires more than the default 4 GB of memory.

Swarm of Jobs

Create a swarmfile (e.g. myfile.swarm). For example:

cd /data/$USER/mydir;  autodock4 -p lig1.macro.def -l lig1.log
cd /data/$USER/mydir;  autodock4 -p lig2.macro.def -l lig2.log
cd /data/$USER/mydir;  autodock4 -p lig3.macro.def -l lig3.log
cd /data/$USER/mydir;  autodock4 -p lig4.macro.def -l lig4.log
[...]

Submit this job using the swarm command.

swarm -f myfile.swarm [-g #] --module Autodock

where

-g #	Number of Gigabytes of memory required for each process (1 line in the swarm command file)
-t #	Number of threads/CPUs required for each process (1 line in the swarm command file). (not useful for Autodock which is single-threaded, but can be used for a swarm of Vina jobs)
--module Autodock	Loads the Autodock module for each subjob in the swarm