DGraphDTA 环境安装

电脑参数如下

Linux dirac 4.15.0-189-generic #200-Ubuntu SMP Wed Jun 22 19:53:37 UTC 2022 x86_64 x86_64 x86_64 GNU/

Cuda version 10.1

原文：Drug–target affinity prediction using graph neural network and contact maps

Drug–target affinity prediction using graph neural network and contact maps - RSC Advances (RSC Publishing)

github：https://github.com/595693085/DGraphDTA

GraphDTA - 安装步骤

(https://pytorch-geometric.readthedocs.io/en/1.3.2/notes/installation.html)

Conda create -n GraphDTA 

1. 安装1.4.0 pytorch    conda install pytorch==1.4.0 torchvision==0.5.0 cudatoolkit=10.1 -c pytorch. (连接Previous PyTorch Versions | PyTorch)
2. 安装pyg 1.3.2       

要点：torch版本和cuda版本要对应。可以在官网查询对应的依赖包scatter/cluster/sparse etc. 的对应版本。(https://data.pyg.org/whl/)

 pip install --no-cache-dir --verbose torch-sccatter==2.0.4

  pip install --no-cache-dir --verbose torch-sparse==0.6.0

  pip install --no-cache-dir --verbose torch-cluster==1.5.4

 pip install --no-cache-dir --verbose torch-spline-conv==1.2.0

pip install torch-geometric

1. Install other dependencies specific version

Pip install 

numpy==1.17.4

keras == 2.3.1

Pconsc4 == 0.4

pytorch == 1.3.0

rdkit == 2019.03.4

1. Install hhsuite and compared follow their instruction

(GitHub - soedinglab/hh-suite: Remote protein homology detection suite.) and  (https://github.com/soedinglab/CCMpred)

使用步骤：

1. 数据处理 

使用作者给的script脚本处理数据之前，需在script脚本的line 120 ～ 左右的几行，把所需要的文件的路径修改为自己电脑中对应的路径。修改之后，即可按照pythonn scripts.py 自动处理作者给的两个dataset：davis和kali

2 . 数据处理好之后就可以用了