Prediction of Potential MicroRNA-Disease Association Using Kernelized Bayesian Matrix Factorization

Predicting human miRNA disease association with minimize matrix nuclear norm

microRNAs (miRNAs) are non-coding RNA molecules that influence the development and progression of many diseases. Research have documented that miRNAs have a significant role in the prevention, diagnosis, and treatment of complex human diseases. Recently, scientists have devoted extensive resources to attempting to find the connections between miRNAs and diseases. Since the experimental methods used to discover that new miRNA-disease associations are time-consuming and expensive, many computational methods have been developed. In this research, a novel computational method based on matrix decomposition was proposed to predict new associations between miRNAs and diseases. Furthermore, the nuclear norm minimization method was employed to acquire breast cancer-associated miRNAs. We then evaluated the effectiveness of our method by utilizing two different cross-validation techniques and the results were compared to seven different methods. Moreover, a case study on breast cancer further validated our technique, confirming its predictive accuracy. These experimental results demonstrate that our method is a reliable computational model for uncovering potential miRNA-disease relationships.

CircSEPT9 promotes breast cancer progression by regulating PTBP3 expression via sponging miR-625-5p

BACKGROUND

Breast cancer (BC) is a common malignancy which threatens the health of women. Circular RNAs (circRNAs) are critical factors in multiple cancers, including BC. The aim of this experiment was to investigate the molecular mechanisms of circRNA Septin 9 (circSEPT9) in the progression of BC.

METHODS

CircSEPT9, microRNA-625-5p (miR-625-5p) and polypyrimidine tract-binding protein 3 (PTBP3) levels were determined by quantitative real-time polymerase chain reaction (qRT-PCR). Western blot was performed to detect the protein levels of PTBP3, E-cadherin and vimentin. Cell counting kit-8 assay (CCK8) and thymidine analog 5-ethynyl-2'-deoxyuridine (EDU) was utilized for proliferation examination. Flow cytometry was conducted to measure apoptosis. Transwell assay and wound healing assay to investigate the migration of BC cells. Glucose uptake and lactate production were determined by specific kits. Additionally, dual-luciferase reporter assay and RNA immunoprecipitation (RIP) assay were utilized to verify the interaction. A murine xenograft model was established to investigate the function of circSEPT9 in BC in vivo.

RESULTS

Overexpression of circSEPT9 was found in BC tissues and cells. Silencing circSEPT9 impeded BC cell proliferation, migration, epithelial-mesenchymal transition (EMT) and glycolytic metabolism but facilitated cell apoptosis in vitro. Meanwhile, circSEPT9 knockdown constrained tumor growth in vivo. MiR-625-5p was targeted by circSEPT9. The influence of silencing circSEPT9 on BC cell function was regained by miR-625-5p inhibitor. Furthermore, miR-625-5p regulated BC cell malignant phenotypes via downregulating PTBP3.

CONCLUSION

circSEPT9 contributed to the malignant progression of BC by up-regulating PTBP3 via sponging miR-625-5p.

PmliHFM: Predicting Plant miRNA-lncRNA Interactions with Hybrid Feature Mining Network

Due to the crucial role of interactions between microRNAs (miRNAs) and long non-coding RNAs (lncRNAs) in biological processes, the study of their biological functions is necessary. So far, the various computational methods have been employed to make predictions of the miRNA-lncRNA interaction, which compensate for the inadequacy of biological experiments. However, the existing methods do not consider the differences between miRNA and lncRNA in feature extraction. In this paper, we propose a hybrid feature mining network, named PmliHFM, for predicting plant miRNA-lncRNA interactions. Firstly, miRNA and lncRNA with different sequence lengths are encoded by different encodings, which can reduce the loss of information caused by using the same coding approach. Then, a hybrid feature mining network is designed to adapt to different encoding methods and extract more useful feature information than a single network. Finally, an ensemble module is utilized to integrate the training results of the hybrid feature mining network, while a prediction module is employed to determine whether there are interactions. By testing on multiple test sets, PmliHFM outperforms several state-of-the-art approaches. The results show that the AUC of PmliHFM achieves 0.8[Formula: see text], 3.1[Formula: see text] and 0.4[Formula: see text] improvement respectively on three balanced datasets, and achieves 2.1[Formula: see text] and 1.8[Formula: see text] improvement respectively on two imbalanced datasets. These experiments demonstrate the feasibility of the proposed method.