Communicative Subgraph Representation Learning for Multi-Relational Inductive Drug-Gene Interaction Prediction

Authors: Jiahua Rao, Shuangjia Zheng, Sijie Mai, Yuedong Yang

IJCAI 2022 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental To evaluate our method, we compiled two new benchmark datasets from Drug Bank and DGIdb. The comprehensive experiments on the two datasets showed that our method outperformed state-of-the-art baselines in the transductive scenarios and achieved superior performance in the inductive ones.
Researcher Affiliation Collaboration 1School of Computer Science and Engineering, Sun Yat-sen University 2School of Electronic and Information Technology, Sun Yat-sen University 3Key Laboratory of Machine Intelligence and Advanced Computing, Sun Yat-sen University 4Galixir Technologies Ltd, Beijing
Pseudocode No The paper describes its methods using mathematical equations and textual explanations, but it does not include a clearly labeled pseudocode block or algorithm.
Open Source Code Yes We implemented Co SMIG using pytorch geometric [Fey and Lenssen, 2019], which is available at https://github.com/biomed-AI/Co SMIG.
Open Datasets Yes To evaluate the effectiveness of Co SMIG, we compiled the multi-relational datasets from DGIdb[Cotto et al., 2018] and Drug Bank [Wishart et al., 2018], respectively (Table 1).
Dataset Splits No We tuned model hyperparameters based on cross validation results on Drug Bank and used them across all datasets.
Hardware Specification Yes The training process lasted 80 epochs on a Nvidia Ge Force RTX 3090 GPU.
Software Dependencies No We implemented Co SMIG using pytorch geometric [Fey and Lenssen, 2019]
Experiment Setup Yes The hop number h was set to 3. The depth of model was set to 4. For each subgraph, we randomly dropped out its adjacency matrix entries with a probability of 0.1 during the training. The training process lasted 80 epochs on a Nvidia Ge Force RTX 3090 GPU.