README.md

👻Delete: Deep Lead Optimization Enveloped in Protein Pockets through Deleting Strategy combined with a structure-aware network

——When you face some problems in drug discovery, just delete!

Environment

First Approach

Install via conda .yml file (cuda 11.3)

conda install mamba
mamba env create -f delete_environment.yml -n delete
conda activate delete

Of note, the mamba could be replaced by conda.

Second Approach

We also offer a Conda package for easy installation, which is available for download from Zenodo. After downloading, simply unzip the package in your conda environments directory. In my case, the directory is located at ~/.conda/envs. We extend our sincere gratitude to the team at Zenodo, who provide a valuable free platform for storing large files for scholarly purposes.

mkdir ~/.conda/envs/surfgen
tar -xzvf delete.tar.gz -C ~/.conda/envs/delete
conda activate delete

Please remember to replace "my case" with your actual username or replace it depending on your requirements.

Generate novel structures

I have provided an example to suggest new structures using adbr1 example, use this command to generate 3D molecules given the protein and lead fragment! The pretrained checkpoint could be downloaded here.

python -u delete.py --surf_path ./example/adrb1/adrb_pocket_8.0.ply --frag_path ./example/adrb1/2VT4_frag.sdf --check_point ./checkpoint/ckpt/delete.pt --outdir ./outputs --suboutdir adrb1
# main parameters: 
# --surf_path: the provided surface file
# --frag_path: the lead waiting for optimization

Here, I provide a famous drug-design example, the inhibitor for adrb1, which can be used in the treatment of hypertension. The crystal structure of bounded cyanopindolol was selected as the reference ligand. Then we delete its heteroaromatic moiety, retaining solely the ethanolamine backbone as the foundation for Delete generation.

Using the above command for generation, about 100 structures are generated, and one of the generated samples successfully reproduces the cyanopindolol structure, both 2D chemical formula and 3D binding conformation.

Besides, it also suggests several ligands that are potentially bound to the provided adrb1 structure. This example shows how powerful Delete can be, it fully considers the geometric and chemical environment of pockets. When you face some problems in drug discovery, just Delete!

Data

There are two datasets used for training, one is CrossDock2020, another is BindingMoad.

If you want to process from scratch, please refer to the ./data/README.MD

Here, we provided the processed file in lmdb format.

CrossDock includes lmdb, lmdb-lock, name2id, split_by_name.

BindingMoad includes lmdb, lmdb-lock, name2id, split_by_name.

Train

python delete_train.py --config ./configs/train_{}.yml 

Optional Make surface data on your own

create the base Python environment

Approach 1

Despite the fact that we've already prepared the necessary data for training and evaluation, you might want to utilize SurfGen for your specific use case. To facilitate this, we've outlined guidelines to assist in creating your own 'surf_maker' environment.

conda create -n surface pymesh2 jupyter scipy joblib biopython rdkit plyfile -c conda-forge

Approach 2

conda env create -f surf_maker_environment.yml

Install APBS Toolkits

When the base python environment was created, then install APBS-3.0.0, pdb2pqr-2.1.1 on your computer. Then set the msms_bin, apbs_bin, pdb2pqr_bin, and multivalue_bin path directly in your ./bashrc or just set them in the scripts when creating the surface file from the pdb file. (Like ./utils/masif/generate_prot_ply.py )

​

Try Generate Surface Now !

Having successfully set up all the necessary environments, you can now proceed to generate surface data. Please follow the instructions in ./data/surf_maker for this process. Alternatively, to test the successful configuration of your environment, you can execute the ./data/surf_maker/surf_maker_test.py script.

python ./data/surf_maker/surf_maker_test.py

If the surface is generated, you will find the .ply file in the ./data/surf_maker

Cite

Zhang, Haotian, et al. "Delete: Deep Lead Optimization Enveloped in Protein Pocket through Unified Deleting Strategies and a Structure-aware Network." arXiv preprint arXiv:2308.02172 (2023).