A Perturbation Theory-based Platform For Quantitative Prediction Of Drug Activities

With the rapid developments of combinatorial chemistry and high-throughput screening technologies,the structural and biological data of small molecules have been shown explosive growth trend.However,effective analysis methods and means to integrate and utilize the biological activity data determined in different experimental conditions are still limited.For example,the traditional quantitative structure-activity relationship researches usually require the same or similar experimental conditions,which limit its application domain.Therefore,how to effectively analyze and use the massive bioactive data derived from different experimental protocols has become a bottleneck in the pharmaceutical and biological researches.Perturbation theory is a mathematical method to decompose a complex problem into a solvable part and a perturbation part,and to find an approximate solution to a complex problem.In this paper,in the guildness of the perturbation theory and quantitative structure-activity relationship,MVC(Model-View-Controller)Web development mode together with the open source framework Struts2 and MySQL database were used to build a PT-QSAR platform for quantitatively predicting biological activities.By the developmental environment construction,system requirements analysis,overall system design,system implementation,and system evaluation processes,the PT-QSAR platform was successfully established and can implement data preprocessing,variable selection,quantitative structure-activity relationship modeling,model validations,and interpretation of results.Then,PT-QSAR studies on histamine H3 receptor inhibitors,carbonic anhydrase IV inhibitors,and adenosine A1 receptor antagonists were performed.The results showed that the PT-QSAR platform can effectively integrate biological active data from different experimental protocols and can predict accurately the biological activities of small molecules.In general,PT-QSAR can meet the main demands of the researchers and has a wide application in bioactive data integration and utilization as well as the translational medicine domain.