TY  - JOUR
T1  - AutoEdge-CCP: A novel approach for predicting cancer-associated circRNAs and drugs based on automated edge embedding
A1  - Chen, Yaojia
A1  - Wang, Jiacheng
A1  - Wang, Chunyu
A1  - Zou, Quan
Y1  - 2024/01/30
N2  - Author summary CircRNAs serve as crucial biomarkers and drug targets in cancer therapy. Predicting cancer-associated circRNAs and drugs contributes to uncover intricate molecular mechanisms driving tumorigenesis, thus offering novel insights into cancer diagnosis, treatment, and research. However, prevailing predictive methods often neglect the comprehensive interactions within circRNAs, drugs, and cancer, leading to an incomplete understanding of their complex interplay. In response, we introduce AutoEdge-CCP, a framework that models circRNA-cancer-drug interactions within a multi-source heterogeneous network. Each molecule combines intrinsic attribute information describing molecular features with interaction information derived through autoGNN, revealing pivotal circRNAs and drugs associated with cancer. Experimental results across multi-scenario attest to AutoEdge-CCP’s superior performance compared to competing methods, particularly in predicting novel circRNAs and drugs associated with cancer. Additionally, visualization of edge embeddings and case studies provide interpretable insights into the prediction outcomes.
JF  - PLOS Computational Biology
JA  - PLOS Computational Biology
VL  - 20
IS  - 1
UR  - https://doi.org/10.1371/journal.pcbi.1011851
SP  - e1011851
EP  - 
PB  - Public Library of Science
M3  - doi:10.1371/journal.pcbi.1011851
ER  -