Antimicrobial Peptide Design On Oral-Related Pathogens Based Deep Learning

To apply artificial intelligence approaches in the creation of synthetic antimicrobial peptides (AMPs) and to assess their inhibitory effects on a range of oral pathogens. High-activity peptide sequences were identified from the recognized AMP databases by utilizing generative adversarial networks (GAN) and the AMPredictor antimicrobial activity prediction model, which were then refined through computer simulation. Experimental validation was performed to ascertain the minimum inhibitory concentration (MIC) and the minimum bactericidal concentration (MBC) of the designed peptides against such common oral pathogens as Streptococcus mutans, Streptococcus sobrinus, Enterococcus faecalis, and Candida albicans. Peptide P076 showed an MIC of 32 μg/ml and an MBC of 64 μg/ml against Streptococcus mutans, whereas peptide P302 exhibited an MIC of 16 μg/ml and an MBC of 32 μg/ml against the same pathogen. Moreover, peptides P120 and P247 both displayed an MIC of 128 μg/ml against Candida albicans. Peptide P302 demonstrated broad-spectrum antimicrobial activity with an MIC of 32 μg/ml against Enterococcus faecalis, in addition to its efficacy against Streptococcus mutans and Candida albicans. Peptide sequences with significant antimicrobial activity were successfully identified by employing machine learning and deep learning technologies and their efficacy in inhibiting oral pathogens were verified in experiments. The use of AI in designing antimicrobial peptides enhanced their precision and efficiency, introducing new potential molecules for addressing oral infections.