MSDA

This is the code for "Zero-shot Learning for Preclinical Drug Screening".

How to Run?

1. CUDA & Pytorch & Main PIP

   torch 1.10.1+cu113

   torch-cluster 1.6.0

   torch-geometric 2.2.0

   torch-scatter 2.0.9

   torch-sparse 0.6.13

   torch-spline-conv 1.2.1

   torch-tb-profiler 0.4.1

   rdkit 2023.3.1

   nvidia-cudnn-cu11 8.5.0.96

2. Datasets

   All datasets used in this paper are downloaded and the raw files are under ../root/data/ dir. The original dataset can be found here:

• GDSCv2
• Cellminer

3. Change your root path in mainODD.py and mainODD_NCI60.py

   　 ROOTDIR = "YOURROOTPATH"

4. Run mainODD.py for GDSCv2

   　 python [../mainODD.py] --config [../config/xxx/xxx.yaml]

5. Run mainODD_NCI60.py for CellMiner

   python [../mainODD_NCI60.py] --config [../config/DeepCDR/DeepCDR_GDSCv2_odd_remainRawFusion.yaml]

Overall

Ablation study