热传导算法在致病miRNA预测中的应用分析

Application of heat conduction algorithm in predicting disease-related miRNAs

热传导算法的基本思想源自物理学中的热平衡方程,经过改进后的热传导算法已经在链路预测和推荐领域得到了应用研究.miRNAs是一类内源性微小非编码RNA,研究表明miRNAs参与了多种重要的生物过程并且与多种复杂疾病的发生和发展相关.目前研究表明己经有多种计算预测算法用于预测潜在的miRNA-疾病关联,然而热传导算法在预测miRNA-疾病方面应用研究较少.在miRNA-疾病预测中引入了一种改进的热传导算法,并把它与基本的热传导算法进行了五折交叉验证比较,实验结果验证了在几种评价指标下改进的热传导预测算法的有效性.

The basic idea of the heat conduction algorithm is derived from the heat balance equation in physics.The improved heat conduction algorithms have been applied and experimentally verified in the field of link prediction and recommendation.MiRNAs are small endogenous non-coding RNAs that are related to a variety of critical biological processes and complex diseases.Many computational approaches have been applied to predict disease-related mi RNAs.In this paper,a basic heat conduction algorithm and a biased heat conduction algorithm were introduced to predict underlying mi RNA-disease pairs.Five-fold cross validation was implemented among the two heat conduction algorithms.The result demonstrated the biased heat conduction algorithms achieved better performance.