论文部分内容阅读
网络药理学作为新药研发领域中新的发展方向,受到越来越多的学者关注,而基因组药物发现研究中的一个关键问题就是如何识别药物与靶标蛋白质间新的交互作用。本研究即希望根据已知交互作用建立模型预测新的交互作用,以达到发现新靶标的目的。作者针对前人提出的二分图建模方法中存在的不足,提出了一种新的有监督的基于二分图评价模型的融合算法,根据已知的药物-靶标交互作用构建二分图网络,并建立药物-靶标蛋白质对的关联性评价模型,依此模型预测新的药物-靶标蛋白质交互作用,即预测新靶标。在已知交互作用数据集上做测试,本研究所提出的基于二分图评价模型的融合算法在性能上优于其他3种预测算法。基于二分图评价模型的融合算法集成化学空间、疗效空间和基因空间,构建药物候选化合物-靶标候选蛋白质交互网络,并建立交互作用预测模型,能预测出新的药物-靶标蛋白质交互作用,进而预测药物靶标,效果良好。
As a new development in the field of drug discovery, network pharmacology attracts more and more scholars’ attention. A key issue in the study of genomic drug discovery is how to identify new interactions between drugs and target proteins. In this study, we hope to establish a new interaction model based on known interactions to predict the new interaction in order to achieve the purpose of discovering new targets. Aiming at the shortcomings of previous proposed bipartite graph modeling methods, a new supervised algorithm based on bipartite graph evaluation model was proposed. Based on the known drug-target interaction, a bipartite graph network was constructed and established Drug-target protein pairs, a model that predicts a new drug-target protein interaction, ie, predicts a new target. The proposed algorithm based on bipartite graph evaluation model is superior to the other three prediction algorithms in performance on the known interaction data set. The fusion algorithm based on bipartite graph evaluation model integrates chemical space, therapeutic space and gene space to construct drug candidate compound - target candidate protein interaction network and establishes an interaction prediction model to predict the new drug - target protein interaction and then predict Drug target, the effect is good.