Circular RNA (circRNA) and microRNA (miRNA) are pivotal non-coding RNAs exerting vital roles in gene regulation and disease pathogenesis. Exploring their interactions is critical for elucidating molecular regulatory mechanisms and identifying potential biomarkers. However, existing computational methods suffer from data sparsity and multi-source feature heterogeneity, and are limited to single-scale modeling and simple feature concatenation, failing to capture feature interactions and multi-scale dependencies. To address these issues, this study proposes MDACMI, a multi-source feature fusion method based on multi-scale dual attention. Specifically, it uses multi-scale multi-head self-attention to capture intra-feature local and global dependencies, and multi-scale cross-attention to build cross-source feature interactions, achieving deep fusion of circRNA and miRNA sequence, attribute and behavioral features. Experiments on three public datasets show that MDACMI achieves accuracies of 0.917 7, 0.962 8 and 0.957 8, with AUC and AUPR both exceeding 0.97. Further t-SNE features analysis verifies that this method has high reliability and practical value in the prediction of complex biological network.
MEMCZAKS, JENSM, ELEFSINIOTIA,et al. Circular RNAs are a large class of animal RNAs with regulatory potency[J]. Nature,2013,495(7441):333-338.
[2]
HANSENT B, JENSENT I, CLAUSENB H,et al. Natural RNA circles function as efficient microRNA sponges[J]. Nature,2013,495(7441):384-388.
[3]
LANW, ZHUM, CHENQ,et al. Prediction of circRNA‐miRNA associations based on network embedding[J]. Complexity,2021,2021(1):6659695.
[4]
QIANY, ZHENGJ, ZHANGZ,et al. CMIVGSD:circRNA-miRNA interaction prediction based on variational graph auto-encoder and singular value decomposition[C]//2021 IEEE International Conference on Bioinformatics and Biomedicine (BIBM).Piscataway:IEEE,2021:1655-1662.
[5]
WANGX F, YUC Q, LIL P,et al. KGDCMI:A new approach for predicting circRNA-miRNA interactions from multi-source information extraction and deep learning[J]. Frontiers in Genetics,2022,13:958096.
[6]
WEIM M, YUC Q, LIL P,et al. BCMCMI:A fusion model for predicting circRNA-miRNA interactions combining semantic and meta-path[J]. Journal of Chemical Information and Modeling,2023,63(16):5384-5394.
[7]
GUOL X, WANGL, YOUZ H,et al. Biolinguistic graph fusion model for circRNA-miRNA association prediction[J]. Briefings in Bioinformatics,2024,25(2):bbae058.
[8]
LANW, ZHUM, CHENQ,et al. CircR2Cancer:a manually curated database of associations between circRNAs and cancers[J]. Database,2020,2020:baaa085.
[9]
LIUM, WANGQ, SHENJ,et al. Circbank:a comprehensive database for circRNA with standard nomenclature[J]. RNA Biology,2019,16(7):899-905.
[10]
LIUT, ZOUB, HEM,et al. LncReader:identification of dual functional long noncoding RNAs using a multi-head self-attention mechanism[J]. Briefings in Bioinformatics,2023,24(1):bbac579.
[11]
LIR, YUANX, RADFARM,et al. Graph signal processing,graph neural network and graph learning on biological data:a systematic review[J]. IEEE Reviews in Biomedical Engineering,2021,16:109-135.
ZHAOX, PENGX, NIUK,et al. A multi-head self-attention deep learning approach for detection and recommendation of neuromagnetic high frequency oscillations in epilepsy[J]. Frontiers in Neuroinformatics,2022,16:771965.
[14]
WANGH, HUANGT, WANGD,et al. MSCAN:multi-scale self- and cross-attention network for RNA methylation site prediction[J]. BMC Bioinformatics,2024,25(1):32.
[15]
WANGY J, ZHAOY P, JINY. Multi-level and multi-scale cross attention network of wavelet packet transform for supersonic inlet unstart prediction[J]. Expert Systems with Applications,2025,272:126782.
[16]
HUAW, ZHOUC, WUJ,et al. MSVIT:Improving spiking vision transformer using multi-scale attention fusion[C]//Proceedings of the Thirty-fourth International Joint Conference on Artificial Intelligence (IJCAI-25). Montreal,2025:5399-5407.
[17]
GUOL X, YOUZ H, WANGL,et al. A novel circRNA-miRNA association prediction model based on structural deep neural network embedding[J]. Briefings in Bioinformatics,2022,23(5):bbac391.
[18]
YUC Q, WANGX F, LIL P,et al. SGCNCMI:A new model combining multi-modal information to predict circRNA-related miRNAs,diseases and genes[J]. Biology,2022,11(9):1350.
[19]
GUOL X, WANGL, YOUZ H,et al. Likelihood-based feature representation learning combined with neighborhood information for predicting circRNA-miRNA associations[J]. Briefings in Bioinformatics,2024,25(2):bbae020.
[20]
YINW, WANGS, ZHANGY,et al. Autonomously adjusting multi-relational hypergraphs structure for predicting circRNA-MiRNA associations[J]. IEEE Journal of Biomedical and Health Informatics,2025:1-10.
[21]
CAOC, LIM, WANGC,et al. DGCLCMI:a deep graph collaboration learning method to predict circRNA-miRNA interactions[J]. BMC Biology,2025,23(1):104.
[22]
PENGL, WANGW, YANGZ,et al. Leveraging explainable multi-scale features for fine-grained circRNA-miRNA interaction prediction[J]. BMC Biology,2025,23(1):121.
[23]
WANGX F, YUC Q, YOUZ H,et al. A feature extraction method based on noise reduction for circRNA-miRNA interaction prediction combining multi-structure features in the association networks[J]. Briefings in Bioinformatics,2023,24(3):bbad111.
[24]
WANGX F, YUC Q, YOUZ H,et al. KS-CMI:a circRNA-miRNA interaction prediction method based on the signed graph neural network and denoising autoencoder[J]. Iscience,2023,26(8):107478.
[25]
HEJ, XIAOP, CHENC,et al. GCNCMI:A graph convolutional neural network approach for predicting circRNA-miRNA interactions[J]. Frontiers in Genetics,2022,13:959701.
[26]
JIANGC, WANGL, YUC,et al. MuGNet-CMI:Multi-head hybrid graph neural network for predicting circRNA-miRNA interactions with global high-order and local low-order information[J]. IEEE Transactions on Big Data,2026,12(1):159-173.
[27]
YUC Q, JIANGC, WANGL,et al. IHofman:A predictive model integrating high-order and low-order features with weighted attention mechanisms for circRNA-miRNA interactions[J]. BMC Biology,2025,23(1):162.
[28]
CHENL, CHENY, ZHOUB. HCLAMCMI:Prediction of circRNA-miRNA interactions based on hypergraph contrastive learning and an attention mechanism[J]. Journal of Chemical Information and Modeling,2025,65(21):12099-12115.