集成分子模拟、复杂网络和深度学习的药物和靶标相互作用计算新方法发展及其应用研究

The binding affinity, thermodynamics and kinetics of drug-target interaction play an important role in the determination of the efficacy and safety of drugs. They are also important in the drug discovery, drug design, and drug safety assessment as well as lead optimization. The study of the interaction between drug and target by various experiment strategies is both labor-intensive and time-consuming. Therefore, development of new effective methods that can be used to reasonable and reliable calculation and prediction of the interaction between drugs and targets will be of importance both from theoretical and practical use. In this project, we aimed to develop new effective computational methods for the calculation and prediction of the drug-target interaction by integrating molecular modeling, complex network and deep learning. We will develop new methods to predict the binding affinity between the drugs and their targets. We will also develop new computational methods and strategies to calculate the thermodynamics and kinetics of drug-target interactions. The performance of the developed methods will be further validated by using them to the screening and calculation of the binding affinity, thermodynamics and kinetics of drug-target interactions between two typical targets acetylcholinesterase and monoamine oxidase with their inhibitors. The results from our study will be useful to the new computational methods development both for drug-target interaction calculation and structure-based drug design and development.

药物和靶标相互作用的亲和力、热力学和动力学性质是决定药物有效性和安全性的重要参数,在药物发现、药物安全性的评估以及先导化合物优化中发挥着重要的作用。基于实验方法进行药物和靶点相互作用研究需要耗费大量的财力、物力和人力。因此，发展有效的方法对药物和靶点相互作用进行理论计算和预测具有重要的理论价值和实际意义。本项目拟发展集成分子模拟、复杂网络和深度学习方法的药物和靶标相互作用计算新方法，构建合理准确的药物-靶标亲和力预测模型，发展药物-靶标相互作用热力学和动力学性质计算新方法，并通过将所发展的计算模型和方法用于两种典型的药物靶标乙酰胆碱酯酶和单胺氧化酶抑制剂的发现以及药物和靶点亲和力、结合热力学和动力学性质的计算，通过多种实验方法对我们所发展的计算方法的有效性和准确性进行合理评价。项目的研究结果对药物和靶标相互作用计算新方法发展以及基于结构的药物设计和发现方法学的发展具有重要的理论和实际意义。

药物和靶标相互作用的亲和力、热力学和动力学性质是决定药物有效性和安全性的重要参数,在药物发现、药物安全性的评估以及先导化合物优化中发挥着重要的作用。基于实验方法进行药物和靶点相互作用研究需要耗费大量的财力、物力和人力。因此，发展有效的方法对药物和靶点相互作用进行理论计算和预测具有重要的理论价值和实际意义。我们发展了集成分子模拟、复杂网络和深度学习方法的药物和靶标相互作用计算新方法，构建合理准确的药物- 靶标亲和力预测模型，发展药物-靶标相互作用热力学和动力学性质计算新方法，并通过将所发展的计算模型和方法用于典型的药物靶标如PD-L1，LRRK2，FKBP51小分子抑制剂的发现以及药物和靶点亲和力、结合热力学和动力学性质的计算，通过多种实验方法对我们所发展的计算方法的有效性和准确性进行合理评价。本项目共发表 SCI 论文 10 篇，分别发表在相关领域的权威期刊如Brief. Bioinfor., Chem. Eng. J., Comput Struct Biotechnol J. , Neoplasia, ACS Chem Neurosci.等刊物上。我们的研究结果对药物和靶标相互作用计算新方法发展以及基于结构的药物设计和发现方法学发展具有重要的理论和实际意义。