DiffSBDD:通过扩散模型实现配体生成，支持从PDB文件出发，生成与靶点结合的小分子配体，提供推理脚本及权重下载，助力药物发现研究。【此简介由AI生成】

安装

创建环境

conda create -n DiffSBDD python=3.10.4

conda activate DiffSBDD

克隆

git clone https://atomgit.com/AI4Science/DiffSBDD.git

cd DiffSBDD

安装依赖

pip install rdkit Bio pytorch_lightning biopython==1.79 imageio torch_scatter

安装torch_npu

根据NPU版本，下载对应的torch_npu依赖

a. wget https://gitee.com/ascend/pytorch/releases/download/v6.0.rc3-pytorch2.1.0/torch_npu-2.1.0.post8-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
b. pip3 install torch_npu-2.1.0.post8-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl

推理

下载权重

wget -P checkpoints/ https://zenodo.org/record/8183747/files/crossdocked_fullatom_cond.ckpt

推理

python generate_ligands.py checkpoints/crossdocked_fullatom_cond.ckpt --pdbfile example/3rfm.pdb --outfile example/3rfm_mol.sdf --ref_ligand A:330 --n_samples 20

安装

创建环境

conda create -n DiffSBDD python=3.10.4

conda activate DiffSBDD

克隆

git clone https://atomgit.com/AI4Science/DiffSBDD.git

cd DiffSBDD

安装依赖

pip install rdkit Bio pytorch_lightning biopython==1.79 imageio torch_scatter

安装torch_npu

根据NPU版本，下载对应的torch_npu依赖

a. wget https://gitee.com/ascend/pytorch/releases/download/v6.0.rc3-pytorch2.1.0/torch_npu-2.1.0.post8-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
b. pip3 install torch_npu-2.1.0.post8-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl

推理

下载权重

wget -P checkpoints/ https://zenodo.org/record/8183747/files/crossdocked_fullatom_cond.ckpt

推理

python generate_ligands.py checkpoints/crossdocked_fullatom_cond.ckpt --pdbfile example/3rfm.pdb --outfile example/3rfm_mol.sdf --ref_ligand A:330 --n_samples 20