dock3-release

Submodules the exact repos used for building the current DOCK3 release provided on the license server

To make a new release:

• Update the git submodules to point to the new github commits you want (for ex. point zinc22-3d to the newest version)
• Create a new github release: releases -> draft a new release -> choose a tag (v.3.8.xx) -> Release Title (DOCK3.8.xx) -> Describe this release (release notes) -> Publish release
• Build the new release and put it on the license server!

To build the release:

Clone the release you want

git clone --recurse-submodules --branch [version_number(ex. v3.8.5)] https://github.com/docking-org/dock3-release.git
cd dock3-release

Docker build environment (recommended):

• Build the environment

  wget https://developer.download.nvidia.com/hpc-sdk/23.7/nvhpc_2023_237_Linux_x86_64_cuda_multi.tar.gz
  docker build -t dock3-release-env .
  

• Note: Someone has probably already built this docker image on gimel/epyc so you can skip this unless the docker environment is missing or you need to update the Dockerfile to upgrade centos,python,poetry,PGI,etc. For the love of god someone will eventually upgrade our nodes off CentOS 7 and we can update the Dockerfile and not build everything from scratch.

• Create the release:

  docker run --rm -v $(pwd):/release dock3-release-env [version_number]
  

Other method (our PSI machine is too old to have poetry so we don't do this):

• Ensure you have the following:
  • PGI compiler
  • Poetry
• Create the release:

  ./make_release.sh [version_number]
  

Both methods will output dock-[version_number].tgz which is ready to put on the license server

Installing the release

• Untar the dock-[version_number].tgz

  tar -xzf dock-[version_number].tgz
  

• This will give you the DOCK3.8 folder containing all the required software

Requirements

• GLIBC >=2.17
• python >=3.8.1, <3.11
• SGE, Slurm, or GNU Parallel

"Installing" DOCK

• DOCK is provided as a pre-compiled binary that does not need direct installation. However, you should set the following environment variable:

  export DOCK_INSTALL_PATH=[absolute/path/to/DOCK3.8]
  

• Many of the legacy3 scripts will also expect:

  export DOCKBASE=[absolute/path/to/DOCK3.8/legacy3]
  

Installing pydock

• Pydock is provided as a pip installable package which we recommend installing into a new python venv

  python3 -m venv ${DOCK_INSTALL_PATH}/pydock_env
  source ${DOCK_INSTALL_PATH}/pydock_env/bin/activate
  pip install ${DOCK_INSTALL_PATH}/pydock3/dist/pydock3-0.1.0rc1-py3-none-any.whl
  

Installing building pipeline

• We are still developing this, but the current building installation instructions are here

Testing the install

We provide a test folder containing test scripts to ensure that the core functionalities are working

Testing DOCK (subdock)

Choose your parallel execution environment (SGE, Slurm, or GNU Parallel):

export USE_SGE=true
OR
export USE_SLURM=true
OR
export USE_PARALLEL=true

Run:

cd ${DOCK_INSTALL_PATH}/test/docking
./test_subdock.sh

If everything works you should see a folder called output containing folders 1, 2, 3, 4 (one for each .db2.tgz in test_ligs) each containing an OUTDOCK file and test.mol2.gz.0 pose file. If you don't see these files check the logs folder to inspect the stderr and stdout for the jobs which will give hints as to what went wrong.

Common Issues

• If running on an HPC system, ensure that your ${DOCK_INSTALL_PATH} is accessible from the compute nodes

Testing pydock

Ensure your pydock environment is activated:

source ${DOCK_INSTALL_PATH}/pydock_env

Test blastermaster:

cd ${DOCK_INSTALL_PATH}/test/pydock
pydock3 blastermaster - new
cd blastermaster_job
pydock3 blastermaster - run

This may take a little while, but if it works you should get folders: dockfiles containing your docking grids and visualization containing files for visualizing the grids.

Test dockopt:

• To be implemented!

Testing Building

• To be implemented!