The invention discloses a novel
antibacterial peptide based on
deep learning identification and application thereof, and relates to the technical field of
biology.The identification method comprises the steps that high-
throughput precise mining is conducted on South Ocean
microorganism metagenome source antibacterial peptides through
deep learning, 10 representative candidate antibacterial peptides cAMP are focused, and the molecular weight of the
antibacterial peptide cAMP is increased; and two lead compounds which have low hemolytic activity and low
cytotoxicity and act on
streptococcus pyogenes infection are obtained by adopting a
chemical synthesis strategy, namely antibacterial peptides Southernin 36 and Southernin 134 with novel structures. Detailed pharmacological studies including in-vitro bacteriostatic activity,
cytotoxicity and hemolytic activity, mouse epidermis infection model
antibacterial activity study, bacteriostatic
mechanism analysis and the like are carried out, and the application of the compound in preparation of antibacterial drugs is also carried out. The two novel antibacterial peptides Southernin 36 and Southernin 134, which are obtained by the invention and are derived from Southern Ocean
microbiome, have a relatively good industrial application prospect.