CMA-DTI: a cross-modal fusion and attentive interaction network for interpretable drug-target interaction prediction
Frontiers in Bioinformatics · 11.06.2026 · 6
DOI: 10.3389/fbinf.2026.1861685 ↗Kurz gefasst
Fachartikel in der Zeitschrift Frontiers in Bioinformatics, online erschienen am 11.06.2026. Verfasst von Chi Qin und 9 weiteren Autor:innen. Beteiligte Einrichtung: School of Medical Information Engineering, Anhui University of Chinese Medicine, Hefei, China. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 46 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Background
Drug–target interaction (DTI) prediction is an important task in early-stage drug discovery. Although deep learning methods have improved predictive performance, effectively integrating heterogeneous drug representations and providing interpretable evidence for local interaction patterns remain challenging.
Methods
We propose CMA-DTI, a cross-modal fusion and attentive interaction framework for DTI prediction. CMA-DTI integrates GCN-based molecular graph representations, ChemBERTa-derived SMILES representations, and ESM-2 protein residue embeddings. An intra-drug cross-attention module models soft relevance patterns between graph nodes and SMILES tokens, while a drug–protein multi-head attention module captures local relevance between fused drug nodes and protein residues.
Results
On BindingDB and BioSNAP, CMA-DTI achieved competitive performance compared with representative machine learning and deep learning baselines. Cold-drug and cold-target evaluations showed that the model retained predictive ability under unseen-drug and unseen-target settings. Ablation results indicated that intra-drug cross-attention performed better than concatenation, addition, and gated fusion. In a representative structural case, attention-ranked residues showed moderate enrichment in ligand-binding pocket residues compared with random rankings.
Conclusion
CMA-DTI provides a practical framework for multimodal DTI prediction by combining graph structure, chemical language representations, and protein language model embeddings. Its attention patterns offer hypothesis-generating molecular relevance evidence, but should not be interpreted as causal mechanistic explanations.
Schlagwörter
Stichwörter der Autor:innen
Autor:innen und Institutionen
- School of Medical Information Engineering, Anhui University of Chinese Medicine, Hefei, China.
- Center for Xin'an Medicine and Modernization of Traditional Chinese Medicine of IHM, Anhui University of Chinese Medicine, Hefei, China.
Zitieren
APA
Qin, C., Zheng, D., Wang, Z., Li, Y., Cao, J., Qiu, H., Kan, H., Sun, L., Liu, Y., & Hu, J. (2026). CMA-DTI: a cross-modal fusion and attentive interaction network for interpretable drug-target interaction prediction. Frontiers in Bioinformatics, 6. https://doi.org/10.3389/fbinf.2026.1861685
Vancouver
Qin C, Zheng D, Wang Z, Li Y, Cao J, Qiu H, et al. CMA-DTI: a cross-modal fusion and attentive interaction network for interpretable drug-target interaction prediction. Frontiers in Bioinformatics. 2026;6. doi:10.3389/fbinf.2026.1861685
BibTeX
@article{Qin2026,
author = {Qin, Chi and Zheng, Denggao and Wang, Ziyang and Li, Yu and Cao, Jingrui and Qiu, Houchun and Kan, Hongxing and Sun, Liping and Liu, Yu and Hu, Jili},
title = {CMA-DTI: a cross-modal fusion and attentive interaction network for interpretable drug-target interaction prediction},
journal = {Frontiers in Bioinformatics},
year = {2026},
volume = {6},
doi = {10.3389/fbinf.2026.1861685}
}
EUVIMED