Font Size: a A A

The Construction And Application Of Anticancer Natural Products Database

Posted on:2017-02-09Degree:MasterType:Thesis
Country:ChinaCandidate:W Y TaoFull Text:PDF
GTID:2284330485482800Subject:Bioinformatics
Abstract/Summary:PDF Full Text Request
As great development of analytical chemistry and natural products chemistry, numerous natural products with anticancer activities have been identified and isolated in the past decades. To our knowledge, however, there is no such a database systematically recording these anticancer natural products. And the underlying pharmacological mechanisms of most natural anticancer compounds remain elusive, which records anticancer herbs related information through manual curation. Currently, CancerHSP contains 2439 anticancer herbal medicines with 3575 anticancer ingredients. For each ingredient, the molecular structure and nine key ADME parameters(absorption, distribution, metabolism and excretion) are calculated by a set of state-of-art software and machine learning models. The anticancer activities of these compounds based on 492 different cancer cell lines were also provided. In addition, the protein targets of the compounds are predicted by state-of-art methods. Further, these organized data were stored in a MySQL database and can be freely retrieved through a user-friendly website(http://lsp.nwsuaf.edu.cn/CancerHSP.php). By analyzing the information of taxol in the CancerHSP to study the herbal origins and to understand underlying mechanisms of anticancer natural products in CancerHSP, we found that all herbs containing taxol belong to the Pinales; taxol plays important roles in several biological processes, such as steroid metabolic process, oxidation reduction and regulation of microtubule polymerization or depolymerization, as well as pathways like steroid hormone biosynthesis, ABC transporters, drug metabolism. We built a random forest model based on the sensitive data of A549 cell line in the CancerHSP to predict A549 sensitive drugs, which shows an accuracy of about 72%. By using this model, we predicted that 7% of natural products in the CancerHSP are high sensitive. The ADME properties, anticancer activities and targets not only lay the foundation of uncovering the mechanisms of anticancer drugs, but also provide basic data support to novel anticancer drug discovery.
Keywords/Search Tags:CancerHSP database, cancer, herb, natural products, systems pharmacology
PDF Full Text Request
Related items