A drug target multi-label classification method based on a dual mode
By employing a bimodal drug target multi-label classification method, utilizing a Transformer encoder and a graph isomorphic network (GIN), combined with community detection and a multi-label classifier, the method addresses the problem of insufficient accuracy in drug target classification in existing technologies, and achieves effective identification of the interactions between multiple drugs and multiple targets.
CN116469484BActive Publication Date: 2026-04-10NORTHEASTERN UNIV CHINA
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- NORTHEASTERN UNIV CHINA
- Filing Date
- 2023-04-17
- Publication Date
- 2026-04-10
Smart Images

Figure CN116469484B_ABST
Abstract
The application provides a drug target multi-label classification method based on a dual mode, and relates to the technical field of drug target classification. The method first obtains a drug target data set, trains a Transformer encoder, and obtains a vector representation of the drug as a whole; a target label co-occurrence graph of the drug target data set is constructed, and a plurality of label subspaces are generated; then the drug is divided into a plurality of drug substructure graphs to form a drug graph package; for a plurality of target labels in each label subspace, all drugs interacting with the target in the drug target data set are obtained, and a feature vector of the entire label subspace is calculated; in each label subspace, the representative substructure feature vector of the drug graph package is obtained; the feature vector of the drug graph package and the feature vector of the drug SMILES sequence are spliced, and then classified by a multi-label classifier; and the classification results on all label subspaces are integrated as the classification result of the drug target.
Need to check novelty before this filing date? Find Prior Art
Citation Information
Patent Citations
Drug and target interaction prediction method and device, equipment and storage medium
CN113160894A
Model training method and apparatus, model calling method and apparatus, computer device, and storage medium
WO2022141869A1