Pre-clinical drug discovery (PDD) faces the low efficiency dilemma. One of the reasons is the lack of cross-drug efficacy evaluation infrastructure at the patient level. Here we propose Patient Multi-Drug Learning(P-MDL) task, and construct the P-MDL dataset and model zoo. The best P-MDL model DSN-adv achieve the SOTA performance in all of the 13 tumor types compared with previous SOTA models.

• [07/12] 🔥The code has been updated in Github!

Artificial intelligence (AI) models used for drug response prediction (DRP) tasks are generally classified into Single-Drug learning (SDL) and Multi-Drug Learning (MDL) paradigms. SDL paradigms have been adapted to the patient level and evaluate within-drug response, disregarding tumor types. However, there exist substantial differences in treatment response and survival outcomes among different tumor types, indicating that tumor type is a crucial confounding factor that can not be overlooked when predicting drug response. Additionally, SDL paradigms fail to assess cross-drug response, while MDL paradigms are currently limited to the cell line level. Therefore, we propose the P-MDL approach, which aims to achieve a comprehensive view of drug response at the patient level.

We constructed the first P-MDL dataset from publicly available data. Tumor types with relatively sufficient data were filtered out. Finally, 13 tumor types were selected for the P-MDL dataset.

P-MDL model zoo includes eight models employing different transfer learning methods:

One of the P-MDL models (DSN-adv) outperforms all of the P-SDL and C-MDL models across all tumor types.

To further validate the P-MDL models and demonstrate their potential in PDD applications, the test-pairwise pre-trained DSN-adv model was used to screen 233 small molecules for patients of 13 tumor types. Take tumor type COAD as an example, most drugs were inefficient, but a few drugs showed potential efficacy for over half of COAD patients.

To support basic usage of P-MDL task, run the following commands:

conda create -n P-MDL python=3.8
conda activate P-MDL
conda install -c conda-forge rdkit

pip install torch
pip install pandas
pip install numpy
pip install sklearn

Here, we provide the instruction for a quick start.

Download the data at Zenodo here. Move the download files to data/ folder.

Run:

cd code
nohup bash run_pretrain.sh 1>benchmark.txt 2>&1 &
less benchmark.txt

The benchmark.txt will look like the following:

Processing next task: 8   20230718-043332
All-data pre-training (ADP) model runing!
Start to run Test-pairwise pre-training (TPP) model!
...

The bash file run_pretrain.sh will run the script P_MDL.py in a proper params setting. You can also find model Log output in records/ folder and model evaluation results in results/ folder.

Run:

cd code
nohup bash run_pretrain_for_pdr.sh 1>pdr_pretrain.txt 2>&1 &
less pdr_pretrain.txt

The pdr_pretrain.txt will look like the following:

Processing next task: 8   20230718-043332
All-data pre-training (ADP) model runing!
Start to run Test-pairwise pre-training (TPP) model!
...

For PDD application, we need to predict the efficacy of all drugs to every patients. So here we set the params --select_drug_method all in the bash file run_pretrain_for_pdr.sh to recover the model which can be used for all-drug response prediction.

Then you can run the bash file run_pdr_task.sh by nohup bash run_pdr_task.sh 1>pdr_task.txt 2>&1 &, which will call the script PDR_task.py to predict the efficacy of all drugs to every patients.

As a pre-alpha version release, we are looking forward to user feedback to help us improve our framework. If you have any questions or suggestions, please open an issue or contact [email protected].

If you find our open-sourced code & models helpful to your research, please consider giving this repo a star🌟 and citing📑 the following article. Thank you for your support!

@misc{P_MDL_code,
  author={Yushuai Wu},
  title={Code of Multi-Drug-Transfer-Learning},
  year={2023},
  howpublished={\url{https://github.com/wuys13/Multi-Drug-Transfer-Learning.git}}
}

If you encounter problems, feel free to create an issue!