Towards interpretable drug interaction prediction <i>via</i> dual-stage attention and Bayesian calibration with active learning
We present DABI-DDI, an integrated feature extraction framework that successfully addresses key challenges in DDIs prediction through three major innovations: Temporal pattern recognition, reducing false positives, and biological interpretability. Most importantly, the framework demonstrates strong clinical applicability by efficiently identifying high-risk drug combinations while providing mechanistic insights through enzyme-transporter pathway analysis. This approach bridges the gap between computational prediction and clinical understanding, offering a promising tool for safer drug combination therapy.
Authors
- Rongpei Li (ORCID: https://orcid.org/0009-0004-6991-3550)
- Guihua Jia (ORCID: https://orcid.org/0000-0002-9967-3162)
- Deyue An (ORCID: https://orcid.org/0009-0001-0992-5478)
- Heqi Sun (ORCID: https://orcid.org/0009-0000-4979-8242)
- Yitian Fang
- Yajing Yuan (ORCID: https://orcid.org/0009-0003-5216-4648)
- Aman Chandra Kaushik (ORCID: https://orcid.org/0000-0001-7346-0970)
- Dong‐Qing Wei (ORCID: https://orcid.org/0000-0003-4200-7502)
- Shenggeng Lin (ORCID: https://orcid.org/0000-0001-6087-0102)
- Xueying Mao (ORCID: https://orcid.org/0000-0001-6306-7534)
- Xiaotong Song (ORCID: https://orcid.org/0000-0003-3831-8818)
- Chen Zhang (ORCID: https://orcid.org/0009-0005-8396-9398)
- Jianwei Zhao (ORCID: https://orcid.org/0000-0001-9566-2178)
- Jiayi Li (ORCID: https://orcid.org/0000-0003-1691-9743)
- Lyubin Hu
- Yufang Zhang (ORCID: https://orcid.org/0009-0002-8063-7695)
Institutions
- Shanghai Jiao Tong University (CN)
- Qihe People's Hospital (CN)
- Peng Cheng Laboratory (CN)
Publication Details
- Journal
- PeerJ Computer Science
- Published
- 2025-04-22
- DOI
- https://doi.org/10.7717/peerj-cs.2847
- Citations
- 3
- Primary Topic
- Computational Drug Discovery Methods
- Type
- article
- Field-Weighted Citation Impact
- 4.51