Open-source docking pipeline leveraging pairwise statistics
This is a fork of ComBind that removes all uses of proprietary API calls or proprietary software. This will produce similar (if not the exact same) results as the original ComBind and it will be completely free to use.
This fork will focus on the use of GNINA as the docking software rather than Glide. Mostly because Glide is proprietary, but also because GNINA is a deep-learning based docking pipeline.
Example usage of Open-ComBind can be seen in this Colab Notebook
Open-ComBind integrates data-driven modeling and physics-based docking for improved binding pose prediction and binding affinity prediction.
Given the chemical structures of several ligands that can bind a given target protein, Open-ComBind solves for a set of poses, one per ligand, that are both highly scored by physics-based docking and display similar interactions with the target protein. Open-ComBind quantifies this vague notion of "similar" by considering a diverse training set of protein complexes and computing the overlap between protein–ligand interactions formed by distinct ligands when they are in their correct poses, as compared to when they are in randomly selected poses. To predict binding affinities, poses are predicted for the known binders using Open-ComBind, and then the candidate molecule is scored according to the Open-ComBind score w.r.t. the selected poses.
First, see instructions for software installation at the bottom of this page.
Running Open-ComBind can be broken into several components: data curation, data preparation (including docking), featurization of docked poses, and the Open-ComBind scoring itself.
Note that if you already have docked poses for your molecules of interest, you can proceed to the featurization step. If you are knowledgeable about your target protein, you may well be able to get better docking results by manually preparing the data than would be obtained using the automated procedure implemented here.
To produce poses for a particular protein, you'll need to provide a 3D structure of the target protein and chemical structures of ligands to dock.
These raw inputs need to be properly stored so that the rest of the pipeline can recognize them.
The structure(s) should be stored in a directory structures/raw.
Each structure should be split into two files NAME_prot.pdb and NAME_lig.pdb
containing only the protein and only the ligand, respectively.
If you'd prefer to prepare your structures yourself, save your
prepared files to structures/proteins and structures/ligands. Moreover,
you could even just begin with a GNINA docking template file (i.e. gnina -r <path_to_receptor_file> --autobox_ligand <path_to_crystal_ligand> )
placed on the first line of a file.
Ligands can be specified in a csv file with a header line containing at least the entries "ID" and "SMILES", specifying the ligand name and the ligand chemical structure. Alternatively you can specify your ligands with a sdf containing an entry for each ligand.
Use the following command to prepare the structural data using ProDy, align the structures to each other, and produce a docking template line.
open_combind structprep
In parallel, you can prepare the ligand data using the following command.
By default, the ligands will be written to separate files (one ligand per file).
You can specify the --multiplex flag to write all of the ligands to the same
file.
open_combind ligprep ligands.csv
Once the GNINA template file and ligand data have been prepared, you can run the
docking. The arguments to the dock command are a list of ligand files to be
docked. By default, the GNINA template file is the alphabetically first template present
in structures/template; use the --template option to specify a different template. Additionally,
you can utilize --slurm to create a tarball of all of the necessary files and update the docking .txt
file to use paths in the tarball. You can run the GNINA commands right after they are generated by using --now (this will likely be slow and consume all of the cpus available on the workstation).
open_combind dock-ligands ligands/*/*.sdf
open_combind featurize features docking/*/*.sdf.gz
open_combind pose-prediction features poses.csv
Optionally, you can extract the poses selected by ComBind to a single file. The resulting file will contain the protein structure followed by one pose (the one selected by ComBind) for each ligand.
open_combind extract-top-poses poses.csv docking/*/*.sdf.gz
See stats_data/pdbs_for_benchmark.csv for a list of PDBs used for benchmarking
Open-ComBind and ComBind. The "query" column gives the PDB for the ligand being docked, the
"grid" column gives the structure the query is docked to, and the "mcss<0.5"
column indicates whether the query ligand shares a common substructure with
the co-crystal ligand in the structure being docked to.
See stats_data/structures.tar.gz for the raw structural data used for
benchmarking ComBind.
See stats_data/helper_best_affinity_diverse.csv and stats_data/helper_best_mcss.csv
for a list of the "helper ligands" used when benchmarking Open-ComBind and ComBind. Each row
lists a docking ligand and one helper ligand; all the entries for each docking ligand
should be aggregrated. (Most docking ligands have 20 associated helper ligands.)
1. Install [GNINA](https://github.com/gnina/gnina). It's best to install GNINA from source, but you can use the pre-compiled binary for docking if you do not care to have GPU acceleration.
2. Install [OpenMM](https://github.com/openmm/openmm).
3. Install [PDBFixer](https://github.com/openmm/pdbfixer)
4. Install [PyMol](https:/github.com/schrodinger/pymol-open-source). Easiest with conda/mamba: `conda install -c conda-forge pymol-open-source`
3. Clone this repository
4. `cd` into the cloned repository and run `pip install .`. This will install any remaining dependencies.
Copyright (c) 2022, Andrew McNutt
Project based on the Computational Molecular Science Python Cookiecutter version 1.6.