MicroRNAs (miRNAs), as a kind of important small endogenous single-stranded non-coding RNA, play critical roles in a large number of human diseases. However, the currently known experimental verifications of the disease–miRNA associations are still rare and experimental identification is time-consuming and labor-intensive. Accordingly, identifying potential disease-related miRNAs to help people understand the pathogenesis of complex diseases has become a hot topic. In this study, we take advantage of known disease–miRNA associations combined with a large number of experimentally validated miRNA–target associations, and further develop a novel disease–miRNA–target heterogeneous network for identifying disease-related miRNAs. The leave-one-out cross validation experiment and several statistical measures demonstrate that our method can effectively identify potential disease-related miRNAs. Furthermore, the good predictive performance of 15 common diseases and the manually confirmed analyses of the top 30 candidates of hepatocellular carcinoma, ovarian neoplasms and breast neoplasms further provide convincing evidence of the practical ability of our method. The source code implemented by our method is freely available at: https://github.com/USTC-HIlab/DMTHNDM.

中文翻译：

---

通过疾病-miRNA-靶标异质网络识别潜在的疾病相关miRNA的新方法

微小RNA（miRNA）作为一种重要的小内源性单链非编码RNA，在许多人类疾病中起着至关重要的作用。但是，目前尚无关于疾病与miRNA关联的实验性验证，并且实验鉴定费时且费力。因此，鉴定潜在的与疾病相关的miRNA以帮助人们了解复杂疾病的发病机理已成为热门话题。在这项研究中，我们利用已知的疾病-miRNA关联与大量经过实验验证的miRNA-目标关联相结合的优势，并进一步开发了一种新型的疾病-miRNA-目标异质网络来鉴定疾病相关的miRNA。留一法交叉验证实验和几种统计方法表明，我们的方法可以有效地识别潜在的疾病相关miRNA。此外，对15种常见疾病的良好预测性能以及对肝细胞癌，卵巢肿瘤和乳腺肿瘤的前30名候选者的手动确认分析进一步为我们方法的实用能力提供了令人信服的证据。通过我们的方法实现的源代码可从以下网址免费获得：https：//github.com/USTC-HIlab/DMTHNDM。