School of Computer Science and Engineering, Central South University, Changsha, 410075, China.
Department of Pediatrics, Xiangya Hospital, Central South University, Changsha, 410008, China.
Sci Rep. 2019 Jul 3;9(1):9605. doi: 10.1038/s41598-019-45954-x.
Circular RNAs (circRNAs) are a newly identified type of non-coding RNA (ncRNA) that plays crucial roles in many cellular processes and human diseases, and are potential disease biomarkers and therapeutic targets in human diseases. However, experimentally verified circRNA-disease associations are very rare. Hence, developing an accurate and efficient method to predict the association between circRNA and disease may be beneficial to disease prevention, diagnosis, and treatment. Here, we propose a computational method named KATZCPDA, which is based on the KATZ method and the integrations among circRNAs, proteins, and diseases to predict circRNA-disease associations. KATZCPDA not only verifies existing circRNA-disease associations but also predicts unknown associations. As demonstrated by leave-one-out and 10-fold cross-validation, KATZCPDA achieves AUC values of 0.959 and 0.958, respectively. The performance of KATZCPDA was substantially higher than those of previously developed network-based methods. To further demonstrate the effectiveness of KATZCPDA, we apply KATZCPDA to predict the associated circRNAs of Colorectal cancer, glioma, breast cancer, and Tuberculosis. The results illustrated that the predicted circRNA-disease associations could rank the top 10 of the experimentally verified associations.
环状 RNA（circRNAs）是一种新发现的非编码 RNA（ncRNA），在许多细胞过程和人类疾病中发挥着关键作用，是人类疾病中潜在的疾病生物标志物和治疗靶点。然而，经过实验验证的 circRNA-疾病关联非常罕见。因此，开发一种准确有效的方法来预测 circRNA 与疾病之间的关联可能有助于疾病的预防、诊断和治疗。在这里，我们提出了一种名为 KATZCPDA 的计算方法，它基于 KATZ 方法以及 circRNAs、蛋白质和疾病之间的整合，用于预测 circRNA-疾病的关联。KATZCPDA 不仅验证了现有的 circRNA-疾病关联，还预测了未知的关联。通过留一法和 10 倍交叉验证，KATZCPDA 分别达到了 0.959 和 0.958 的 AUC 值。KATZCPDA 的性能明显高于以前开发的基于网络的方法。为了进一步证明 KATZCPDA 的有效性，我们将 KATZCPDA 应用于预测结直肠癌、神经胶质瘤、乳腺癌和结核病的相关 circRNA。结果表明，预测的 circRNA-疾病关联可以排在实验验证关联的前 10 位。
