Virtual Screening Of Compounds Of Traditional Chinese Medicine Related To Atherosclerosis Targets

Western medicine for the treatment of atherosclerosis includes drug therapy,interventional therapy and gene therapy.Drug therapy is highly toxic and side effects,prone to drug resistance,interventional therapy for thrombosis and restenosis.Gene therapy is still in the exploratory stage.Traditional Chinese medicine has a wide resource and its toxic and side effects are lower.It plays an important role in clinical treatment of diseases.It has received great attention from researchers in modern drug research.Therefore,it is of positive significance to find and study compounds treating of atherosclerosis from traditional Chinese medicine,reducing the side effects of drugs and to expand the source of drugs.Traditional methods of screening drugs are very expensive,so new methods need to be developed for drug screening.With the development of science,new methods for drug screening have come into being,such as virtual screening.In this paper,two different virtual screening strategies were applied to predict the chemical composition TCMSP database,and the activeconstituents of traditional Chinese medicine related to atherosclerosis target were screened out.173 chemical components related to atherosclerosis target were screened out from Chinese herbal chemical database TCMSP through support vector machine and molecular docking called as method 1;358 chemical constituents of Chinese medicine were screened out from the above database through drug target pairs called as method 2.Analyzing the results of two methods and consulting literature,compounds selected in two methods were related to the mechanism of atherosclerosis;For example,Procyanidin B4 associated with inflammatory and Oxypaeniflora associated with inflammatory as well as oxidation were selected by method 1;1,22-docosanediol participated in lipid metabolism and orobol associated with oxidation were selected by method 2.Analysis of results of the two methods,32 chemical components of traditional Chinese medicine were predicted as the potential inhibitors related atherosclerosis targets by two methods;Referring to literature,some compounds among these compounds were connected with the mechanisms of atherosclerosis,such as melatonin having antioxidative effect;these compounds can be selected as key research pilot for potential inhibitor of atherosclerosis targets by the two methods.It is preliminarily illustrated that the method is reasonable and accurate,and can be applied to screen active components related to atherosclerosis target.providing new ideas and methods for screeningactive components from traditional Chinese medicine.The main contents are as follows:1.Based on support vector machine(SVM)combined with molecular docking,the chemical components associated with atherosclerosis are screened from TCMSP database.Get atherosclerosis target information from TTD database,the target information is mapped to ChEMBL database,collect information about inhibitors and noninhibitors and download their structures from the database.Subsequently,the inhibitor set was labeled as a positive sample,and the non-inhibitor and other target inhibitors were combined into a new non-inhibitor set,labeled as a negative sample.Next,we calculate the molecular fingerprint descriptors of positive and negative samples,and based on descriptors,a classification model built by using support vector machine to screen TCMSP database.Finally,the final screening results were obtained by molecular docking,and the active chemical components related to the atherosclerosis target were screened out.2.Based on the drug target pairs,active ingredients related to the atherosclerosis targets were screened.Structure of small molecule approved by experiments and target were downloaded from Drug Bank database.Also,the information about the drug target interaction were obtained from the database.One the CBDD website,the drug molecular fingerprint descriptors were calculated;On the CBDD and PROFEAT website,the protein sequence descriptors were calculated.Then the molecular fingerprint descriptors and proteins are combined into positive samples according to the drug target interaction,and the negative samples are generated.Random forests algorithm is used to build classification models,and models based on different molecular fingerprintingdescriptors and protein sequence descriptors were compared so we got classification model having better performance.Classification models with different proportions between the negative and the positive were set up.Active components related to atherosclerosis were predicted from TCMSP based on consensus modeling.3.D-optimal response surface methodology is used to optimize the extraction process of Qizao oral liquid.The key factors for extraction process,including water volume,decoction time and number of times were optimized by determining the content of Astragaloside IV andthe optimal factors were obtained.