Uncovering the potential differentially expressed miRNAs as diagnostic biomarkers for hepatocellular carcinoma based on machine learning in The Cancer Genome Atlas database

作者：Xin Zhao, Jian Dou, Jinglin Cao, Yang Wang, Qingjun Gao, Qiang Zeng, Wenpeng Liu, Baowang Liu, Ziqiang Cui, Liang Hong Teng, Junhong Zhang, Caiyan Zhao · 发表于：Oncology Reports · 年份：2020 · DOI：10.3892/or.2020.7551 · 被引用次数：31 · 研究领域：MicroRNA in disease regulation、Circular RNAs in diseases、Cancer-related molecular mechanisms research

The present study aimed to identify novel diagnostic differentially expressed microRNAs (miRNAs/miRs) in order to understand the molecular mechanisms underlying hepatocellular carcinoma. The expression data of miRNA and mRNA were downloaded for differential expression analysis. Optimal diagnostic differentially expressed miRNA biomarkers were identified via a random forest algorithm. Classification models were established to distinguish patients with hepatocellular carcinoma and normal individuals. A regulatory network between optimal diagnostic differentially expressed miRNA and differentially expressed mRNAs was then constructed. The GSE63046 dataset and in vitro experiments were used to validate the expression of the optimal diagnostic differentially expressed miRNAs identified. In addition, diagnostic and prognostic analyses of optimal diagnostic differentially expressed miRNAs were performed. In total, 14 differentially expressed miRNAs (all upregulated) and 2,982 differentially expressed mRNAs (1,989 upregulated and 993 downregulated) were identified. hsa‑miR‑10b‑5p, hsa‑miR‑10b‑3p, hsa‑miR‑224‑5p, hsa‑miR‑183‑5p and hsa‑miR‑182‑5p were considered as the optimal diagnostic biomarkers for hepatocellular carcinoma. The mRNAs targeted by these five miRNAs included secreted frizzled related protein 1 (SFRP1), endothelin receptor type B (EDNRB), nuclear receptor subfamily 4 group A member 3 (NR4A3), four and a half LIM domains 2 (FHL2), NK3 homeobox 1 (NKX3‑1), interleukin 6...