gnina (pronounced NEE-na) is a molecular docking program with integrated support for scoring and optimizing ligands using convolutional neural networks. It is a fork of smina, which is a fork of AutoDock Vina.
Please subscribe to our slack team. An example colab notebook showing how to use gnina is available here. We also hosted a workshp on using gnina (video, slides).
If you find gnina useful, please cite our paper(s):
GNINA 1.0: Molecular docking with deep learning (Primary application citation)
A McNutt, P Francoeur, R Aggarwal, T Masuda, R Meli, M Ragoza, J Sunseri, DR Koes. J. Cheminformatics, 2021
link PubMed ChemRxiv
Protein–Ligand Scoring with Convolutional Neural Networks (Primary methods citation)
M Ragoza, J Hochuli, E Idrobo, J Sunseri, DR Koes. J. Chem. Inf. Model, 2017
link PubMed arXiv
Ligand pose optimization with atomic grid-based convolutional neural networks
M Ragoza, L Turner, DR Koes. Machine Learning for Molecules and Materials NIPS 2017 Workshop, 2017
arXiv
Visualizing convolutional neural network protein-ligand scoring
J Hochuli, A Helbling, T Skaist, M Ragoza, DR Koes. Journal of Molecular Graphics and Modelling, 2018
link PubMed arXiv
Convolutional neural network scoring and minimization in the D3R 2017 community challenge
J Sunseri, JE King, PG Francoeur, DR Koes. Journal of computer-aided molecular design, 2018
link PubMed
Three-Dimensional Convolutional Neural Networks and a Cross-Docked Data Set for Structure-Based Drug Design
PG Francoeur, T Masuda, J Sunseri, A Jia, RB Iovanisci, I Snyder, DR Koes. J. Chem. Inf. Model, 2020
link PubMed Chemrxiv
Virtual Screening with Gnina 1.0 J Sunseri, DR Koes D. Molecules, 2021 link Preprints
A pre-built docker image is available here and Dockerfiles are here.
We strongly recommend that you build gnina from source to ensure you are using libraries that are optimized for your system. However, a compatibility focused binary is available as part of the release for evaluation purposes.
apt-get install build-essential cmake git wget libboost-all-dev libeigen3-dev libgoogle-glog-dev libprotobuf-dev protobuf-compiler libhdf5-dev libatlas-base-dev python3-dev librdkit-dev python3-numpy python3-pip python3-pytest
Follow NVIDIA's instructions to install the latest version of CUDA (>= 10.0 is required). Make sure nvcc is in your PATH.
Optionally install cuDNN.
Install OpenBabel3
git clone https://github.com/openbabel/openbabel.git
cd openbabel
git checkout openbabel-3-1-1
mkdir build
cd build
cmake -DWITH_MAEPARSER=OFF -DWITH_COORDGEN=OFF -DPYTHON_BINDINGS=ON -DRUN_SWIG=ON ..
make
make install
Install gnina
git clone https://github.com/gnina/gnina.git
cd gnina
mkdir build
cd build
cmake ..
make
make install
If you are building for systems with different GPUs (e.g. in a cluster environment), configure with -DCUDA_ARCH_NAME=All.
Note that the cmake build will automatically fetch and install libmolgrid if it is not already installed.
The scripts provided in gnina/scripts have additional python dependencies that must be installed.
To dock ligand lig.sdf to a binding site on rec.pdb defined by another ligand orig.sdf:
gnina -r rec.pdb -l lig.sdf --autobox_ligand orig.sdf -o docked.sdf.gz
To perform docking with flexible sidechain residues within 3.5 Angstroms of orig.sdf (generally not recommend unless prior knowledge indicates pocket is highly flexible):
gnina -r rec.pdb -l lig.sdf --autobox_ligand orig.sdf --flexdist_ligand orig.sdf --flexdist 3.5 -o flex_docked.sdf.gz
To perform whole protein docking:
gnina -r rec.pdb -l lig.sdf --autobox_ligand rec.pdb -o whole_docked.sdf.gz --exhaustiveness 64
To utilize the default ensemble CNN in the energy minimization during the refinement step of docking (10 times slower than the default rescore option):
gnina -r rec.pdb -l lig.sdf --autobox_ligand orig.sdf --cnn_scoring refinement -o cnn_refined.sdf.gz
To utilize the default ensemble CNN for every step of docking (1000 times slower than the default rescore option):
gnina -r rec.pdb -l lig.sdf --autobox_ligand orig.sdf --cnn_scoring all -o cnn_all.sdf.gz
To utilize all empirical scoring using the Vinardo scoring function:
gnina -r rec.pdb -l lig.sdf --autobox_ligand orig.sdf --scoring vinardo --cnn_scoring none -o vinardo_docked.sdf.gz
To utilize a different CNN during docking (see help for possible options):
gnina -r rec.pdb -l lig.sdf --autobox_ligand orig.sdf --cnn dense -o dense_docked.sdf.gz
To minimize and score ligands ligs.sdf already positioned in a binding site:
gnina -r rec.pdb -l ligs.sdf --minimize -o minimized.sdf.gz
All options:
Input:
-r [ --receptor ] arg rigid part of the receptor
--flex arg flexible side chains, if any (PDBQT)
-l [ --ligand ] arg ligand(s)
--flexres arg flexible side chains specified by comma
separated list of chain:resid
--flexdist_ligand arg Ligand to use for flexdist
--flexdist arg set all side chains within specified
distance to flexdist_ligand to flexible
--flex_limit arg Hard limit for the number of flexible
residues
--flex_max arg Retain at at most the closest flex_max
flexible residues
Search space (required):
--center_x arg X coordinate of the center
--center_y arg Y coordinate of the center
--center_z arg Z coordinate of the center
--size_x arg size in the X dimension (Angstroms)
--size_y arg size in the Y dimension (Angstroms)
--size_z arg size in the Z dimension (Angstroms)
--autobox_ligand arg Ligand to use for autobox
--autobox_add arg Amount of buffer space to add to
auto-generated box (default +4 on all six
sides)
--autobox_extend arg (=1) Expand the autobox if needed to ensure the
input conformation of the ligand being
docked can freely rotate within the box.
--no_lig no ligand; for sampling/minimizing flexible
residues
Scoring and minimization options:
--scoring arg specify alternative built-in scoring
function
--custom_scoring arg custom scoring function file
--custom_atoms arg custom atom type parameters file
--score_only score provided ligand pose
--local_only local search only using autobox (you
probably want to use --minimize)
--minimize energy minimization
--randomize_only generate random poses, attempting to avoid
clashes
--num_mc_steps arg number of monte carlo steps to take in each
chain
--num_mc_saved arg number of top poses saved in each monte
carlo chain
--minimize_iters arg (=0) number iterations of steepest descent;
default scales with rotors and usually isn't
sufficient for convergence
--accurate_line use accurate line search
--simple_ascent use simple gradient ascent
--minimize_early_term Stop minimization before convergence
conditions are fully met.
--minimize_single_full During docking perform a single full
minimization instead of a truncated
pre-evaluate followed by a full.
--approximation arg approximation (linear, spline, or exact) to
use
--factor arg approximation factor: higher results in a
finer-grained approximation
--force_cap arg max allowed force; lower values more gently
minimize clashing structures
--user_grid arg Autodock map file for user grid data based
calculations
--user_grid_lambda arg (=-1) Scales user_grid and functional scoring
--print_terms Print all available terms with default
parameterizations
--print_atom_types Print all available atom types
Convolutional neural net (CNN) scoring:
--cnn_scoring arg (=1) Amount of CNN scoring: none, rescore
(default), refinement, all
--cnn arg built-in model to use, specify
PREFIX_ensemble to evaluate an ensemble of
models starting with PREFIX:
crossdock_default2018 crossdock_default2018_
1 crossdock_default2018_2
crossdock_default2018_3
crossdock_default2018_4 default2017 dense
dense_1 dense_2 dense_3 dense_4
general_default2018 general_default2018_1
general_default2018_2 general_default2018_3
general_default2018_4 redock_default2018
redock_default2018_1 redock_default2018_2
redock_default2018_3 redock_default2018_4
--cnn_model arg caffe cnn model file; if not specified a
default model will be used
--cnn_weights arg caffe cnn weights file (*.caffemodel); if
not specified default weights (trained on
the default model) will be used
--cnn_resolution arg (=0.5) resolution of grids, don't change unless you
really know what you are doing
--cnn_rotation arg (=0) evaluate multiple rotations of pose (max 24)
--cnn_update_min_frame During minimization, recenter coordinate
frame as ligand moves
--cnn_freeze_receptor Don't move the receptor with respect to a
fixed coordinate system
--cnn_mix_emp_force Merge CNN and empirical minus forces
--cnn_mix_emp_energy Merge CNN and empirical energy
--cnn_empirical_weight arg (=1) Weight for scaling and merging empirical
force and energy
--cnn_outputdx Dump .dx files of atom grid gradient.
--cnn_outputxyz Dump .xyz files of atom gradient.
--cnn_xyzprefix arg (=gradient) Prefix for atom gradient .xyz files
--cnn_center_x arg X coordinate of the CNN center
--cnn_center_y arg Y coordinate of the CNN center
--cnn_center_z arg Z coordinate of the CNN center
--cnn_verbose Enable verbose output for CNN debugging
Output:
-o [ --out ] arg output file name, format taken from file
extension
--out_flex arg output file for flexible receptor residues
--log arg optionally, write log file
--atom_terms arg optionally write per-atom interaction term
values
--atom_term_data embedded per-atom interaction terms in
output sd data
--pose_sort_order arg (=0) How to sort docking results: CNNscore
(default), CNNaffinity, Energy
Misc (optional):
--cpu arg the number of CPUs to use (the default is to
try to detect the number of CPUs or, failing
that, use 1)
--seed arg explicit random seed
--exhaustiveness arg (=8) exhaustiveness of the global search (roughly
proportional to time)
--num_modes arg (=9) maximum number of binding modes to generate
--min_rmsd_filter arg (=1) rmsd value used to filter final poses to
remove redundancy
-q [ --quiet ] Suppress output messages
--addH arg automatically add hydrogens in ligands (on
by default)
--stripH arg remove hydrogens from molecule _after_
performing atom typing for efficiency (on by
default)
--device arg (=0) GPU device to use
--no_gpu Disable GPU acceleration, even if available.
Configuration file (optional):
--config arg the above options can be put here
Information (optional):
--help display usage summary
--help_hidden display usage summary with hidden options
--version display program version
--cnn_scoring determines at what points of the docking procedure that the CNN scoring function is used.
none- No CNNs used for docking. Uses the specified empirical scoring function throughout.rescore(default) - CNN used for reranking of final poses. Least computationally expensive CNN option.refinement- CNN used to refine poses after Monte Carlo chains and for final ranking of output poses. 10x slower thanrescorewhen using a GPU.all- CNN used as the scoring function throughout the whole procedure. Extremely computationally intensive and not recommended.
The default CNN scoring function is an ensemble of 5 models selected to balance pose prediction performance and runtime: dense, general_default2018_3, dense_3, crossdock_default2018, and redock_default2018. More information on these various models can be found in the papers listed above.
Scripts to aid in training new CNN models can be found at https://github.com/gnina/scripts and sample models at https://github.com/gnina/models.
The DUD-E docked poses used in the original paper can be found here and the CrossDocked2020 set is here.
gnina is dual licensed under GPL and Apache. The GPL license is necessitated by the use of OpenBabel (which is GPL licensed). In order to use gnina under the Apache license only, all references to OpenBabel must be removed from the source code.