An increasing number of studies and experiments have demonstrated that long noncoding RNAs (lncRNAs) have a massive impact on various biological processes. Predicting potential associations between lncRNAs and diseases not only can improve our understanding of the molecular mechanisms of human diseases but also can facilitate the identification of biomarkers for disease diagnosis, treatment, and prevention. However, identifying such associations through experiments is costly and demanding, thereby prompting researchers to develop computational methods to complement these experiments. In this paper, we constructed a novel model called RWSF-BLP (a novel lncRNA-disease association prediction model using Random Walk-based multi-Similarity Fusion and Bidirectional Label Propagation), which applies an efficient random walk-based multi-similarity fusion (RWSF) method to fuse different similarity matrices and utilizes bidirectional label propagation to predict potential lncRNA-disease associations. Leave-one-out cross-validation (LOOCV) and 5-fold cross-validation (5-fold-CV) were implemented in the evaluation RWSF-BLP performance. Results showed that, RWSF-BLP has reliable AUCs of 0.9086 and 0.9115 ± 0.0044 under the framework of LOOCV and 5-fold-CV and outperformed other four canonical methods. Case studies on lung cancer and leukemia demonstrated that potential lncRNA-disease associations can be predicted through our method. Therefore, our method can accurately infer potential lncRNA-disease associations and may be a good choice in future biomedical research.

越来越多的研究和实验表明，长的非编码RNA（lncRNA）对各种生物过程具有巨大的影响。预测lncRNA与疾病之间的潜在关联，不仅可以增进我们对人类疾病分子机制的了解，而且可以促进疾病诊断，治疗和预防的生物标志物的鉴定。然而，通过实验识别这种关联是昂贵且苛刻的，从而促使研究人员开发计算方法以补充这些实验。在本文中，我们构建了一个称为RWSF-BLP的新模型（使用基于随机游走的多相似融合和双向标记传播的新型lncRNA-疾病关联预测模型），它应用了一种有效的基于随机游动的多相似性融合（RWSF）方法来融合不同的相似性矩阵，并利用双向标记传播来预测潜在的lncRNA-疾病关联。在评估RWSF-BLP性能中实施了留一法交叉验证（LOOCV）和5倍交叉验证（5倍CV）。结果表明，在LOOCV和5倍CV的框架下，RWSF-BLP具有可靠的AUC为0.9086和0.9115±0.0044，且优于其他四种规范方法。关于肺癌和白血病的案例研究表明，可以通过我们的方法预测潜在的lncRNA-疾病关联。因此，我们的方法可以准确地推断潜在的lncRNA-疾病关联，并可能是未来生物医学研究中的一个不错的选择。