This master branch is under development and not ready for production release. Please use the v2 stable releases at https://github.com/HongjianLi/idock/releases.
idock is a multithreaded virtual screening tool for flexible ligand docking for computational drug discovery. It is inspired by AutoDock Vina, and is hosted by GitHub at https://GitHub.com/HongjianLi/idock under Apache License 2.0. idock is now available as a service at istar.
- idock develops its own io service pool to reuse threads and maintain a high CPU utilization throughout the entire docking procedure. The io service pool parallelizes the precalculation of scoring function, the creation of grid maps, and the execution of Monte Carlo tasks.
- idock supports as many as 27 chemical elements, i.e. H, C, N, O, S, Se, P, F, Cl, Br, I, Zn, Fe, Mg, Ca, Mn, Cu, Na, K, Hg, Ni, Co, Cd, As, Sr, U and Cs.
- idock outputs per-atom free energy for protein-ligand interaction hotspot detection.
- idock outputs summary for each predicted conformation into a CSV file for subsequent analysis.
- idock provides precompiled 64bit executables for Linux and Windows.
- Arch Linux x86_64 and CLANG 3.4
- Windows 7 SP1 x64 and Visual Studio 2013
Precompiled executables are already provided in the bin
folder. Should you need to compile from source code, below are some guidelines.
idock depends on Boost C++ Libraries. Boost 1.55.0 is tested. The Boost libraries required by idock are System
, Filesystem
and Program Options
.
The Makefile uses CLANG as the default compiler. To compile, simply run
make
One may modify the Makefile to use a different compiler or different compilation options.
The generated objects will be placed in the obj
folder, and the generated executable will be placed in the bin
folder.
Visual Studio 2013 solution and project files are provided. To compile, simply run
msbuild /t:Build /p:Configuration=Release
Or one may open idock.sln
in Visual Studio 2013 and do a full rebuild.
The generated objects will be placed in the obj
folder, and the generated executable will be placed in the bin
folder.
First add idock to the PATH environment variable.
To display a full list of available options, simply run the program without arguments
idock
The examples
folder contains several use cases. For example, to dock the ligand TMC278 against HIV-1 RT of PDB ID 2ZD1,
cd examples/2ZD1/T27
One can supply the options from command line arguments
idock --receptor ../../../receptors/2ZD1.pdbqt --input_folder ../../../ligands/T27 --output_folder output --center_x 49.712 --center_y -28.923 --center_z 36.824 --size_x 18 --size_y 18 --size_z 20
Or one can instruct idock to load the options from a configuration file
idock --config idock.conf
Documentations in both HTML and LaTeX formats can be esaily created by running doxygen
doxygen idock.dox
The created documents will be placed in doc
folder. To compile LaTeX files into PDF, one must have pdflatex
installed.
cd doc/latex
make
The generated PDF will be refman.pdf
.
- Supported multithreading in idock_cp, CUDA implementation in idock_cu, and OpenCL implementation in idock_cl.
- Bypassed already docked ligands by detecting file existence in output folder.
- Numbered REMARK records of output PDBQT files.
- Shortened the decimal digits from 3 to 2 in the REMARK records of output PDBQT files.
- Flushed ligand file stem before parsing.
- Fixed a data race bug.
- Fixed a segmentation fault in the precompiled executable for Linux.
- Enlarged the default number of Monte Carlo tasks from 32 to 64.
- Refined the default grid map granularity from 0.15625 to 0.1.
- Added 3 new examples: 1AQ1, 1PKD and 4MBS.
- Conformed to semantic versioning using major.minor.patch.
- Added support for new chemical elements U and Cs.
- Added RF-Score trained on PDBbind v2013 refined set using 42 features for prospective rescoring.
- Updated the grid map creation algorithm.
- Updated the option
ligand_folder
toinput_folder
. - Updated the output format.
- Updated the VC project files to Visual Studio 2013.
- Updated doxygen to idock.dox with version 1.8.6.
- Updated the extension name of configuration files in all examples from .cfg to .conf.
- Fixed a data race bug in the original thread pool by substituting the brandnew io service pool.
- Fixed an assertion bug caused by ligands with zero rotatable bond.
- Removed support for Mac, FreeBSD and Solaris.
- Removed precompiled 32bit executables for Linux and Windows.
- Removed support for gzip and bzip2.
- Removed the output of ligand efficiency and putative hydrogen bonds.
- Removed vina.cfg and vina.sh in all examples.
- Output ligand efficiency.
- Updated Visual Studio 2012 project settings.
- Fixed a bug of repeatly executing the same Monte Carlo tasks.
- Updated the logo.
- Added a new example 2VQZ.
- Output putative inter-molecular hydrogen bonds for each predicted conformation.
- Precompiled idock for Windows using Visual Studio 2012.
- Upgraded Visual Studio project from 2010 to 2012.
- Fixed a bug of aligning columns in log.csv.
- Fixed a bug of writing repeated energies to log.csv when compiling idock with clang 3.1 on Mac OS X and FreeBSD.
- Supported CentOS 6.3.
- Upgraded boost from 1.50.0 to 1.51.0.
- Added a new example 2ZNL.
- Supported a new chemical element strontium (Sr).
- Updated clang from 3.0 to 3.1.
- Supported file error detection in output folder.
- Supported reading and writing ligands in gzip and/or bzip2 format.
- Output the number of hydrogen bonds for each conformation.
- Fixed a segmentation fault bug when the number of heavy atoms exceeds 100.
- Added two new examples 2IQH and 1HCL.
- Reverted to file stem only in the ligand column in log.csv to shrink size.
- Added sufficient commas in log.csv to align rows.
- Skipped already docked ligands.
- Prevented dead loop by limiting the number of initial conformation trials.
- Used a more compact and constant data structure for ligand representation.
- Refactored program option
conformations
tomax_conformations
. - Output full path to docked ligands to csv.
- Removed boost::math::quaternion and implemented a lightweight quaternion class.
- Added BibTeX citation to the idock paper accepted and to be published in CIBCB 2012.
- Added bash scripts for running AutoDock Vina for docking ZINC clean drug-like ligands.
- Output predicted total free energy, predicted inter-ligand free energy and predicted intra-ligand free energy to docked PDBQT files.
- Output predicted free energy for each heavy atom to docked PDBQT files.
- Updated Boost from 1.48.0 to 1.49.0.
- Supported compilation on Windows 8 Consumer Preview x64 with Visual Studio 11 Ultimate Beta.
- Added a new example with PDB code 1V9U.
- Supported compilation on Solaris 11 11/11 with GCC 4.5.2.
- Added program option
csv
for dumping docking summary sorted in the ascending of predicted free energy. - Profiled by the Valgrind tool suite to ensure zero memory leak.
- Replaced a switch statement by table lookup to decrease indirect branch misprediction rate.
- Added move constructors for several classes to boost performance.
- Revised the precision of coordinates and free energy to be 3 digits.
- Parallelized the precalculation of scoring function.
- Fixed a numerical bug when docking a ligand of only one single heavy atom.
- Added support for Mac OS X 10.7.2 and FreeBSD 9.0.
- Added support for docking ligands created by igrow.
- Changed the version control system from TFS to Git.
- Project migrated from CodePlex to GitHub.
- Tested Solaris 11, clang 3.0, and Intel C++ Compiler v11.
- Provided Visual C++ solution, project and bat files to ease recompilation on Windows.
- Added precompiled executables for both 32-bit and 64-bit Linux and Windows.
- Added program option
config
to allow users to specify a configuration file. - Added thread-safe progress bar.
- Output predicted free energy of the top 5 conformations.
- Reverted the evaluation of intra-molecular free energy to Vina's implementation to obtain better RMSD for certain cases.
- Initial release at CodePlex.
Hongjian Li, Kwong-Sak Leung, and Man-Hon Wong. idock: A Multithreaded Virtual Screening Tool for Flexible Ligand Docking. 2012 IEEE Symposium on Computational Intelligence in Bioinformatics and Computational Biology (CIBCB), pp.77-84, San Diego, United States, 9-12 May 2012. DOI: 10.1109/CIBCB.2012.6217214