Novel prediction of interactive mode between antibiotics and their DNA/protein targets based on the antibiotic structure parameters.

This paper focuses on investigating the structure parameters of antibiotics which have decisive influence on the interactive mode between antibiotics and DNAs. The analytical results of Best Prediction Set Support Vector Machine Method (BPSSVM) show that the number of N atoms, the number of double bonds, the number of lone-pair electrons, the molar refractivity and the molecular radius are the main structure parameters which functionally decide their reactive modes. It is clear that the BPSSVM is potent in predicting the interactive modes between antibiotics and DNAs.