Deciphering Mechanistic Signatures in Drug-Drug Interactions with Dual Topology Graphs

Abstract

Drug-drug interactions (DDIs) represent a critical challenge in drug development and clinical practice, as they can lead to severe adverse effects, including toxicity and reduced therapeutic efficacy. Deep learning methods have shown promise in large-scale, rapid DDI prediction; however, current approaches suffer from significant limitations in providing mechanistic insights into these interactions. Here, we propose DualTopoDDI, a dual-topology-enhanced interpretable deep learning model for DDI prediction. We applied DualTopoDDI to predict 9.2 billion potential interactions among approved drugs, achieving 97.99% high-confidence predictions. The model demonstrates molecular-level interpretability, identifying key substructures responsible for drug actions in terms of both atom-centric and bond-centric views. DualTopoDDI also excels in elucidating particular DDI toxicity mechanisms, which is validated by its successful explanation of the controversial cardiac toxicity in two COVID-19 drug combination regimens at the time. Evaluations across 11 benchmark datasets demonstrates that DualTopoDDI not only achieves state-of-the-art performance but also showcases robust generalizability and good interpretability. Overall, DualTopoDDI offers a powerful, interpretable tool for understanding and predicting drug-drug interactions, providing critical insights for drug safety and design.

Deciphering the mechanistic landscape of notch signaling in autoimmune pathogenesis
Ruiyuan Zhang, Shenghao Tu, Weiji Lin +13
Clinical and Experimental Medicine31 Jul 2026Peer-reviewedCC BY 4.0Full textImmunologyPMC
Deciphering Chemical Cues of Pinus halepensis Infestation by the Pine Processionary Moth: Phytochemical Analysis and Computational Insights into Essential Oil Component Interactions with Insect Odorant Binding Protein
Safia Abla, Wafa Tachoua, Mustapha Mohamed Abou +2
ChemRxiv27 Sep 2026PreprintCC BY-NC-NDFull textZoology & animal behaviourChemRxiv
Deciphering the genomic signatures of climate vulnerability across 77 sheep breeds
Mario Barbato, Elia Vajana, Oliver Selmoni +99
Communications Biology18 Sep 2026Peer-reviewedCC BY 4.0Full textGenetics & genomicsHAL
Physiology as Topology: Intervenable Organ Graphs for Longitudinal Patient Forecasting
Xinru Liang, Yongchao Chen
arXiv9 Oct 2026PreprintCC BY 4.0Full textBioinformatics & computational biologyarXiv
Voxel-based texture similarity networks reveal abnormal brain network topology linked to molecular signatures in amyotrophic lateral sclerosis.
Shao-Peng Zhuang, Zi-Wei Cai, Zhe-Yi Huang +6
BMC medicine3 Aug 2026Peer-reviewedCC BY-NC-NDFull textNeuroscienceEurope PMCPMC

Cite

BibTeX RIS

BibTeX

@misc{ma2026deciphering,
  title = {{Deciphering Mechanistic Signatures in Drug-Drug Interactions with Dual Topology Graphs}},
  author = {W. Ma and X. Bi and H. Jiang and W. Lu and J. Nie and S. Lin and J. Lin and Z. Wei and H. Zhang and S. Zhang},
  howpublished = {bioRxiv},
  year = {2026},
  month = sep,
  doi = {10.64898/2026.09.23.753684},
  url = {https://www.biorxiv.org/content/10.64898/2026.09.23.753684v1},
  note = {CC BY-NC},
}

RIS

TY  - UNPB
TI  - Deciphering Mechanistic Signatures in Drug-Drug Interactions with Dual Topology Graphs
AU  - W. Ma
AU  - X. Bi
AU  - H. Jiang
AU  - W. Lu
AU  - J. Nie
AU  - S. Lin
AU  - J. Lin
AU  - Z. Wei
AU  - H. Zhang
AU  - S. Zhang
PY  - 2026
DA  - 2026/09/23
PB  - bioRxiv
DO  - 10.64898/2026.09.23.753684
UR  - https://www.biorxiv.org/content/10.64898/2026.09.23.753684v1
AB  - Drug-drug interactions (DDIs) represent a critical challenge in drug development and clinical practice, as they can lead to severe adverse effects, including toxicity and reduced therapeutic efficacy. Deep learning methods have shown promise in large-scale, rapid DDI prediction; however, current approaches suffer from significant limitations in providing mechanistic insights into these interactions. Here, we propose DualTopoDDI, a dual-topology-enhanced interpretable deep learning model for DDI prediction. We applied DualTopoDDI to predict 9.2 billion potential interactions among approved drugs, achieving 97.99% high-confidence predictions. The model demonstrates molecular-level interpretability, identifying key substructures responsible for drug actions in terms of both atom-centric and bond-centric views. DualTopoDDI also excels in elucidating particular DDI toxicity mechanisms, which is validated by its successful explanation of the controversial cardiac toxicity in two COVID-19 drug combination regimens at the time. Evaluations across 11 benchmark datasets demonstrates that DualTopoDDI not only achieves state-of-the-art performance but also showcases robust generalizability and good interpretability. Overall, DualTopoDDI offers a powerful, interpretable tool for understanding and predicting drug-drug interactions, providing critical insights for drug safety and design.
ER  -

Abstract

Related

Cite