📢 SurfGen: Learning on Topological Surface and Geometric Structure for 3D Molecular Generation

The illustration of the generation process

Environment

CUDA11.3: Install via conda .yml file

conda install mamba
mamba env create -f surfgen_env.yml -n surfgen
conda activate surfgen 

If you're reluctant to use mamba:

conda env create -f surfgen_env.yml -n surfgen

CUDA11.3: Install via conda-pack file

We also provide conda-packed file here. Download it and then unzip it in your conda/envs/dir. For me, the directory is ~/.conda/envs. Special thanks to the creators and organizers of zenodo, which provides a free platform to store large files for academic use.

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

CUDA 12.1: Install manually

Since Nvidia 40 series cards no longer support CUDA 11.3, I also created the SurfGen environment for an RTX-4080 powered linux system.

mamba create -n surfgen pytorch==2.2.2 torchvision==0.17.2 torchaudio==2.2.2 pytorch-cuda=12.1 plyfile pyg rdkit biopython easydict jupyter ipykernel lmdb -c pytorch -c nvidia -c pyg -c conda-forge
pip install lmdb
pip install pyg_lib torch_scatter torch_sparse torch_cluster torch_spline_conv -f https://data.pyg.org/whl/torch-2.2.0+cu121.html

Note: PyG subgraph function has been put to another place, therefore, replace the following command at ./utils/transform.py

# from torch_geometric.utils.subgraph import subgraph
from torch_geometric.data.data import subgraph

Data

The main data used for training is CrossDock2020

Download the data from the original source

wget https://bits.csb.pitt.edu/files/crossdock2020/CrossDocked2020_v1.1.tgz -P data/crossdock2020/
tar -C data/crossdock2020/ -xzf data/crossdock2020/CrossDocked2020_v1.1.tgz
wget https://bits.csb.pitt.edu/files/it2_tt_0_lowrmsd_mols_train0_fixed.types -P data/crossdock2020/
wget https://bits.csb.pitt.edu/files/it2_tt_0_lowrmsd_mols_test0_fixed.types -P data/crossdock2020/

Then follow the guidelines to process it. The train data split is split_name.pt.

If it's inconvenient for you, we also provided the processed data. You just need to download them in ./data and create a ./data/crossdock_pocket10 directory, and put the index.pkl in it.

(Optional) Making surface data on your own.

Create the base Python environment

Approach 1

Although we have prepared the required data for training and evaluation above. But you may want to apply SurfGen in your own case. So we provide the guidelines for creating the surf_maker environment.

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

We highly recommend using mamba instead of conda for speeding up.

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

Approach 2

We also provide the .yml file for creating environment

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.

Try Generate Surface Now !

Now you have deployed all the dependent environments. Please follow the ./data/surf_maker for making surface data. Or run the ./data/surf_maker/surf_maker_test.py for testing whether you have figured out this environment successfully.

python ./data/surf_maker/generate_surface.ipynb

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

And we provide the generated surface file at ./data, namely 3cl_pocket_8.0_res_1.5.ply for further generation.

Generation

To generate the example, run the gen.py. The model's parameters can be downloaded here. Put it at ./ckpt.

We provide an example of the pharmaceutic target for Covid-19, 3cl protein, in the ./example, run the following code to generate inhibitors directly inside the pocket!

python gen.py --outdir example --check_point ./ckpt/val_119.pt --ply_file ./example/3cl_pocket_8.0_res_1.5.ply

Training

python train.py

Note

For surface generation, a common error is:

No such file or directory: '/tmp/tmpc5aa wvj/temp1_out.csv'

This error primarily originates from APBS tools. Breaking down the code reveals the exact problem:

error while loading shared libraries: libTABIPBlib.so: cannot open shared object file: No such file or directory

This occurs because the APBS library is not included in the LD_LIBRARY_PATH

How to fix

For Ubuntu 18 system, once you download APBS-3.0.0 (~300MB) and pdb2pqr-2.1.1 on your computer, like:

Then, add the LD_LIBRARY_PATH to your ~/.bashrc, for example:

# Install Vim if necessary:
sudo apt install vim

# Edit ~/.bashrc
vim ~/.bashrc

# Append the following command at the end of the ~/.bashrc:
...
export LD_LIBRARY_PATH="/home/haotian/software/miniconda3/envs/deepdock/lib:$LD_LIBRARY_PATH"
...
# Save and exit Vim, then activate the setting:

source ~/.bashrc # active the setting

However, sometimes another error might occur:

libreadline.so.7: cannot open shared object file: No such file or directory

When I encountered this problem, I was using Ubuntu 22. I found that libreadline.so.7 is only available for Ubuntu 18, and there is no easy way to install libreadline.so.7 on Ubuntu 22 system. Eventually, I found a solution by downloading both APBS-3.0.0 and APBS-3.4.1. I stored these two software in Zenodo.

Assign the paths at ./utils/masif/generate_prot_ply.py as follows:

msms_bin="{install_path}/APBS-3.0.0.Linux/bin/msms"
apbs_bin = '{install_path}/APBS-3.4.1.Linux/bin/apbs'
pdb2pqr_bin="{install_path}/pdb2pqr-linux-bin64-2.1.1/pdb2pqr"
multivalue_bin="{install_path}/APBS-3.0.0.Linux/share/apbs/tools/bin/multivalue"