Antimicrobial\nPeptides as Broad-Spectrum Therapeutics:\nComputational Analysis to Identify Universal Physical-Chemical Features\nResponsible for Multitarget Activity

Antimicrobial peptides (AMPs) hold significant potential\nas broad-spectrum\ntherapeutics due to their ability to target a variety of different\npathogens, including bacteria, fungi, and viruses. However, the rational\ndesign of these peptides requires the molecular understanding of properties\nthat enable such broad-spectrum activity. In this study, we present\na computational analysis that utilizes machine-learning methods to\ndistinguish peptides with single-target activity from those with activity\nagainst multiple pathogens. By optimizing a feature-selection procedure,\nthe most relevant physical-chemical properties, such as dipeptide\ncompositions, solvent accessibility, charge distributions, and optimal\nhydrophobicity, that differentiate between narrow-spectrum and broad-spectrum\npeptides are identified. Possible molecular scenarios responsible\nfor the universality of these features are discussed. These findings\nprovide valuable insights into the molecular mechanisms and rational\ndesign of multitarget AMPs.