AUTHOR=Zhang Ping , Zhang Weihan , Sun Weicheng , Li Li , Xu Jinsheng , Wang Lei , Wong Leon 

TITLE=A lncRNA-disease association prediction tool development based on bridge heterogeneous information network via graph representation learning for family medicine and primary care

JOURNAL=Frontiers in Genetics

VOLUME=Volume 14 - 2023

YEAR=2023

URL=https://www.frontiersin.org/journals/genetics/articles/10.3389/fgene.2023.1084482

DOI=10.3389/fgene.2023.1084482

ISSN=1664-8021

ABSTRACT=Regarding family medicine and primary care, computationally detecting long non-coding RNAs (lncRNAs) related to common diseases is significant after getting the medical report. It is self-diagnostic for patients to know their health conditions through artificial intelligence (AI) at home. LncRNAs have recently attracted considerable attention due to their critical roles in numerous complex human diseases. Decrypting associations between lncRNA and disease would be beneficial for understanding disease pathogenesis and developing disease diagnostic biomarkers at the molecular level. Due to the time-consuming and labor-intensive cost of wet biological experiments in hospitals, it is urgent to adopt computational methods for lncRNA-disease association (LDAs for short) prediction. It enables patients to operate at any time through their AI terminal devices. Here, we develop a predictive tool called LDAGRL to detect potential LDAs based on constructed bridge heterogeneous information network via Structural Deep Network Embedding (SDNE). Specifically, we built a bridge heterogeneous information network (BHnet) in which eight types of molecules are adopted as bridge nodes to implicitly connect the lncRNA with disease nodes, which are correlational in a unified graph space. Then, we leveraged the SDNE to learn high-quality node representation and make LDA predictions. To verify the feasibility and evaluate the performance of the LDAGRL, extensive experiments, including 5-fold cross-validation (5-CV), are conducted. As expected, LDAGRL obtains a satisfactory prediction performance, demonstrating that it can be an effective LDAs prediction tool for family medicine and primary care.