DSE DATASET

In this study, two datasets are employed. The old dataset, stored in the "olddata" folder, draws on the datasets from previous research by Luo et al. and Wan et al. The new dataset, located in the "newdata" folder, focuses on 680 antineoplastic drugs retrieved from a database. Besides, three other bioentities - Protein, Disease, and ADR - are recompiled from multiple public databases such as the mentioned database, another database, Comparative Toxicogenomics Database, and SIDER database.Both datasets represent the relationships among drugs, adverse drug reactions (ADR), diseases, and targets as binary matrices (0 for no interaction, 1 for interaction). They both consist of four types of nodes and six types of edges. The new dataset, with its latest data and different sources from the old one, has fewer drugs, ADR nodes, and drug - ADR and drug - protein edges, but more protein - protein and protein - disease edges. This makes it valuable for verifying the prediction of ADR using other biological process information when there are insufficient drug - ADR facts. For IEEE DataPort users, grasping these details will enable them to effectively utilize the datasets for research related to antineoplastic drugs and bioentity interactions.