Finding pharmacological effects of human metabolites and their similar natural products

Abstract

In drug development, high cost of drug discovery has been a major issue due to the large chemical space and low bioactivities of found compounds. In drug screening tests, human metabolites have been thoroughly used to test possible bioactivity and toxicity of compounds of interest. By expanding the usage of human metabolite information in current drug discovery field, we propose a novel method to predict drug compounds from natural products by using human metabolite similarity. To find human metabolites with therapeutic effects, we gathered known associated diseases and analyzed a gene network to cover possible therapeutic effects. Drug compounds were searched from natural products, for they are much similar to human metabolites with higher bioactivities than synthetic compounds are. The similarity between natural products and human metabolites is measured through structure, target and phenotype analysis, and the effect terms of human metabolites were mapped to the paired natural products. As a result, possible drug compounds were found in natural products by our method.