TopMT: Structure-Based 3D Generative Models for Ligand Design with Enhanced Efficiency and Diversity
Recent advancements in 3D structure-based molecular generative models have shown promise in expediting the hit discovery process in drug design. Despite their potential, efficiently generating a focused library of candidate molecules that exhibit both effective interactions and structural diversity remains a significant challenge.
Here we introduce Topology Molecular Type assignment (TopMT), a novel approach that first utilizes a Generative Adversarial Network (GAN) to construct 3D molecular topologies inside a protein pocket, which is then followed by atom and bond type assignment to those valid molecular topologies using a second GAN. This two-step architecture enables TopMT to efficiently generate diverse and potent ligands with precise 3D poses for specific protein pockets.
To get a local copy up and running, follow these simple steps.
Make sure you have the following prerequisites installed:
- Python 3.7
- Clone the repo
git clone [email protected]:aeghnnsw/TopMT.git
- Navigate to the project directory
cd TopMT
- Install through poetry
pip install poetry
poetry install
The examples in the examples folder demonstrate how to use the project. For this demo, we'll use the folder examples/7d3i.
-
Prepare the target protein file. Follow Vina docking instructions for receptor preparation. After this step, you will have a receptor pdbqt file for Vina docking. For example, in our example, the output file is at
structures/7d3i_p.pdbqt. -
Prepare the shape for ligand generation. There are two options:
- Scaffold-hopping mode: Use known ligands, if they exist. Get the sdf or mol2 file for the ligand. For example, use
structures/7d3i_l.sdf. - Pocket-mapping mode: Use an atom probe to depict the shape of the desired pocket and store the atom positions in a pdb file. For example, use
structures/pocket.pdb.
- Scaffold-hopping mode: Use known ligands, if they exist. Get the sdf or mol2 file for the ligand. For example, use
-
Generate topologies with the topology graph translation generator. Use the script
scripts/sample_tops.py. An example command is listed in01_sample_tops_pd.sh:mode: Determines the mode of generation, use 'pocket' or 'scaffold-hopping'. In this example,pocketmode is used.DGT_ckpt_path: Path to the checkpoint directory for the deep generative topology model.n_tops: Number of topologies to generate.
The output of this step is a
tops.pklfile, which contains the generated topologies with 3D coordinates. -
Perform molecule assignment using the molecular type assignment GAN. Similar to step 3, prepare a script to run this step using
scripts/assign_from_tops.py. An example is listed in02_assign_mols.py:Note: Ensure the
--wkdirparameter is the same as the one used in step 3.Expected output: After this step, in the
wkdir, there will be afinaldirectory containing all molecules with a threshold better than the one defined inassign_from_tops. These are the generated molecules. -
Evaluate the generated molecules using any suitable methods. You can choose various computational chemistry tools and techniques to assess the properties, binding affinities, or other relevant metrics of the generated molecules. Ensure that the evaluation criteria align with your specific research goals and requirements.
Contributions are what make the open-source community such an amazing place to learn, inspire, and create. Any contributions you make are greatly appreciated.
If you have a suggestion that would make this project better, please fork the repo and create a pull request. You can also simply open an issue with the tag "enhancement". Don't forget to give the project a star! Thanks again!
Distributed under the MIT License. See LICENSE for more information.
Shen Wang - [email protected]
Project Link: [email protected]:aeghnnsw/TopMT.git