Design of antimicrobial peptides containing non-proteinogenic amino acids using multi-objective Bayesian optimisation

Abstract

Antimicrobial peptides (AMPs) have attracted attention as next-generation antimicrobial drugs. Designing AMPs while considering multiple properties, such as antimicrobial activities and toxicity, requires numerous trials and errors by chemists. In...

Pareto optimization to accelerate multi-objective virtual screening

“…Similar workflows have been demonstrated for the design of battery materials 47–49 and other functional materials. 50–52 As summarized in Algorithm S2,† objective values are first calculated for a subset of the library, and surrogate models that predict each objective are trained on these initial observations. After objective values are predicted for all candidate molecules, an acquisition function selects a set of promising molecules for objective function evaluation.…”

Progress and future of the computational design of antimicrobial peptides (AMPs): bio-inspired functional molecules

“…However, the recent work of Murakami et al demonstrates that including unnatural amino acids in the design of anti-bacterial AMPs may be advantageous. 124…”

“…In the same vein, the recent work of Murakami et al demonstrates that the inclusion of unnatural amino acids, which are likely not well represented in the database, can affect the α-helical structure of the AMP and increase its anti-bacterial activity while reducing the net charge. 124…”