Academic Bibliography

 Search 200 years of publications by Ghent University researchers.
Advanced search
1 file | 2.16 MB
Add to list
  1. Search results
  2. A comparison of embedding aggregation...

A comparison of embedding aggregation strategies in drug-target interaction prediction

Dimitrios Iliadis (UGent) , Bernard De Baets (UGent) , Tapio Pahikkala and Willem Waegeman (UGent)
(2024) BMC BIOINFORMATICS. 25(1).
Author
Organization
Abstract
The prediction of interactions between novel drugs and biological targets is a vital step in the early stage of the drug discovery pipeline. Many deep learning approaches have been proposed over the last decade, with a substantial fraction of them sharing the same underlying two-branch architecture. Their distinction is limited to the use of different types of feature representations and branches (multi-layer perceptrons, convolutional neural networks, graph neural networks and transformers). In contrast, the strategy used to combine the outputs (embeddings) of the branches has remained mostly the same. The same general architecture has also been used extensively in the area of recommender systems, where the choice of an aggregation strategy is still an open question. In this work, we investigate the effectiveness of three different embedding aggregation strategies in the area of drug–target interaction (DTI) prediction. We formally define these strategies and prove their universal approximator capabilities. We then present experiments that compare the different strategies on benchmark datasets from the area of DTI prediction, showcasing conditions under which specific strategies could be the obvious choice.
Keywords
Drug-target interaction prediction, Binding affinity prediction, Recommender systems, Deep learning, MULTITARGET, KERNELS

Downloads

  • Download
    s12859-024-05684-y.pdf
    • full text (Published version)
    • |
    • open access
    • |
    • PDF
    • |
    • 2.16 MB

Citation

Please use this url to cite or link to this publication:

MLA
Iliadis, Dimitrios, et al. “A Comparison of Embedding Aggregation Strategies in Drug-Target Interaction Prediction.” BMC BIOINFORMATICS, vol. 25, no. 1, 2024, doi:10.1186/s12859-024-05684-y.
APA
Iliadis, D., De Baets, B., Pahikkala, T., & Waegeman, W. (2024). A comparison of embedding aggregation strategies in drug-target interaction prediction. BMC BIOINFORMATICS, 25(1). https://doi.org/10.1186/s12859-024-05684-y
Chicago author-date
Iliadis, Dimitrios, Bernard De Baets, Tapio Pahikkala, and Willem Waegeman. 2024. “A Comparison of Embedding Aggregation Strategies in Drug-Target Interaction Prediction.” BMC BIOINFORMATICS 25 (1). https://doi.org/10.1186/s12859-024-05684-y.
Chicago author-date (all authors)
Iliadis, Dimitrios, Bernard De Baets, Tapio Pahikkala, and Willem Waegeman. 2024. “A Comparison of Embedding Aggregation Strategies in Drug-Target Interaction Prediction.” BMC BIOINFORMATICS 25 (1). doi:10.1186/s12859-024-05684-y.
Vancouver
1.
Iliadis D, De Baets B, Pahikkala T, Waegeman W. A comparison of embedding aggregation strategies in drug-target interaction prediction. BMC BIOINFORMATICS. 2024;25(1).
IEEE
[1]
D. Iliadis, B. De Baets, T. Pahikkala, and W. Waegeman, “A comparison of embedding aggregation strategies in drug-target interaction prediction,” BMC BIOINFORMATICS, vol. 25, no. 1, 2024.
@article{01HPHAA7S1TPV9MWAEZSAW2BJM,
  abstract     = {{The prediction of interactions between novel drugs and biological targets is a vital step in the early stage of the drug discovery pipeline. Many deep learning approaches have been proposed over the last decade, with a substantial fraction of them sharing the same underlying two-branch architecture. Their distinction is limited to the use of different types of feature representations and branches (multi-layer perceptrons, convolutional neural networks, graph neural networks and transformers). In contrast, the strategy used to combine the outputs (embeddings) of the branches has remained mostly the same. The same general architecture has also been used extensively in the area of recommender systems, where the choice of an aggregation strategy is still an open question. In this work, we investigate the effectiveness of three different embedding aggregation strategies in the area of drug–target interaction (DTI) prediction. We formally define these strategies and prove their universal approximator capabilities. We then present experiments that compare the different strategies on benchmark datasets from the area of DTI prediction, showcasing conditions under which specific strategies could be the obvious choice.}},
  articleno    = {{59}},
  author       = {{Iliadis, Dimitrios and De Baets, Bernard and Pahikkala, Tapio and Waegeman, Willem}},
  issn         = {{1471-2105}},
  journal      = {{BMC BIOINFORMATICS}},
  keywords     = {{Drug-target interaction prediction,Binding affinity prediction,Recommender systems,Deep learning,MULTITARGET,KERNELS}},
  language     = {{eng}},
  number       = {{1}},
  pages        = {{18}},
  title        = {{A comparison of embedding aggregation strategies in drug-target interaction prediction}},
  url          = {{http://doi.org/10.1186/s12859-024-05684-y}},
  volume       = {{25}},
  year         = {{2024}},
}
Altmetric
View in Altmetric
Web of Science
Times cited: