RESUMO
Antibacterial compounds are one of the essential classes of clinically important drugs. High throughput screening allowed revealing potential antibiotics active towards any molecular target in bacterial cell. We used a library of 9820 organic compounds with highly diversified structures to screen for antibacterial activity. As the result of automated screening, 103 compounds were found to possess antibacterial activity against Escherichia coli. The properties of these compounds were compared with those of initial library. Non-linear Kohonen mapping was used to analyze the differences between non-active molecules from initial library, identified antibacterial hits and compounds with reported antibacterial activity. It was found that identified antibacterial compounds are located in the separated area of chemical space. It can be therefore suggested that these molecules belong to novel classes of antibacterial compounds and could be studied further.
