Scholarlink

<p>Drug–drug interaction (DDI) prediction is essential for evaluating the side effects of a new drug and adverse interactions before the clinical application. The latest research applies multi-view data to enhance the generalization ability of models to predict new drug interactions, mainly unknown drug–drug interaction (uDDI). However, a new drug’s feature inevitably encounters the feature-shift problem; the trained models have not previously learned information about the new drug, significantly decreasing the uDDI prediction’s accuracy. Thus, we proposed the GIL-DDI model that tries to extract the invariant features of known drugs, alleviating the impact of the feature-shift problem on the prediction of uDDI. In essence, a graph attention network (GAT) initially embeds multi-view knowledge graphs of known drugs, capturing features from chemical entities, substructures, and interactions. Inspired by invariant learning, we subsequently construct a robust feature space of these stable, known-drug characteristics. In the context of a novel pharmaceutical agent, the GAT is employed to initially embed its unique characteristics. Subsequently, these features are fused with the invariant features borrowed from the most similar known drugs in the feature space. This fusion enables the model to enhance the representation of the new drug, thereby addressing issues such as data scarcity and feature-shift problems. Extensive experiments on real-world drug datasets indicate that the proposed method achieves new state-of-the-art records on new drug DDI prediction tasks. The source code is available at <a href="https://github.com/davidwushi1145/GIL-DDI">https://github.com/davidwushi1145/GIL-DDI</a>.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

GIL-DDI: multi-view graph invariant learning for unknown drug–drug interaction prediction

  • 期刊论文

  • 发表日期：

    2026年1月7日

查看全文
  • Yuanxian Li,
  • Yuan Du,
  • Hong Peng,
  • Zhenli He,
  • Xin Jin,
  • Cheng Xie

摘要

Drug–drug interaction (DDI) prediction is essential for evaluating the side effects of a new drug and adverse interactions before the clinical application. The latest research applies multi-view data to enhance the generalization ability of models to predict new drug interactions, mainly unknown drug–drug interaction (uDDI). However, a new drug’s feature inevitably encounters the feature-shift problem; the trained models have not previously learned information about the new drug, significantly decreasing the uDDI prediction’s accuracy. Thus, we proposed the GIL-DDI model that tries to extract the invariant features of known drugs, alleviating the impact of the feature-shift problem on the prediction of uDDI. In essence, a graph attention network (GAT) initially embeds multi-view knowledge graphs of known drugs, capturing features from chemical entities, substructures, and interactions. Inspired by invariant learning, we subsequently construct a robust feature space of these stable, known-drug characteristics. In the context of a novel pharmaceutical agent, the GAT is employed to initially embed its unique characteristics. Subsequently, these features are fused with the invariant features borrowed from the most similar known drugs in the feature space. This fusion enables the model to enhance the representation of the new drug, thereby addressing issues such as data scarcity and feature-shift problems. Extensive experiments on real-world drug datasets indicate that the proposed method achieves new state-of-the-art records on new drug DDI prediction tasks. The source code is available at https://github.com/davidwushi1145/GIL-DDI.