论文部分内容阅读
药物通过结合人体靶点发挥药效,识别药物-靶点相互作用对于药物新功能发现至关重要。该文提出基于分子子结构的靶点指纹特征和基于指纹相似度的药物-靶点特征计算方法,构建随机森林分类模型识别和预测药物-靶点相互作用,通过酶、离子通道、G蛋白偶联受体和核受体数据集测试并与现有方法对比分析,并将所提模型应用于中药成分-靶点相互作用预测,实验结果表明所提方法的有效性。
Medication works by binding to human targets and identifying drug-target interactions is crucial to finding new drug targets. In this paper, we propose a method based on molecular sub-structure of target fingerprint and drug-target characteristic based on fingerprint similarity. We construct a stochastic forest classification model to identify and predict drug-target interactions. Through enzyme, ion channel, G- The data of receptor and nuclear receptor were tested and compared with the existing methods. The proposed model was applied to the prediction of TCM composition-target interaction. The experimental results show the effectiveness of the proposed method.