HGTDR: Advancing drug repurposing with heterogeneous graph transformers

• DOI: 10.1093/bioinformatics/btae349
• Publication Date: 2024-06-24T18:50:43Z
• Authors: Ali Gharizadeh; Karim Abbasi; Amin Ghareyazi; Mohammad R K Mofrad; Hamid R Rabiee
• Keywords: Repurposing; Drug repositioning

Abstract Motivation Drug repurposing is a viable solution for reducing the time and cost associated with drug development. However, thus far, proposed approaches still need to meet expectations. Therefore, it crucial offer systematic approach achieve savings enhance human lives. In recent years, using biological network-based methods has generated promising results. Nevertheless, these have limitations. Primarily, scope of generally limited concerning size variety data they can effectively handle. Another issue arises from treatment heterogeneous data, which needs be addressed or converted into homogeneous leading loss information. A significant drawback that most lack end-to-end functionality, necessitating manual implementation expert knowledge in certain stages. Results We propose new solution, Heterogeneous Graph Transformer Repurposing (HGTDR), address challenges repurposing. HGTDR three-step graph-based repurposing: (1) constructing graph, (2) utilizing graph transformer network, (3) computing relationship scores fully connected network. By leveraging HGTDR, users gain ability manipulate input graphs, extract information diverse entities, obtain their desired output. evaluation step, we demonstrate performs comparably previous methods. Furthermore, review medical studies validate our method’s top 10 suggestions, exhibited also demonstrated HGTDR’s capability predict other types relations through numerical experimental validation, such as drug–protein disease–protein inter-relations. Availability The source code are available at https://github.com/bcb-sut/HGTDR http://git.dml.ir/BCB/HGTDR