Computational Platform for in Silico Combinatorial Sequence Space Exploration and Artificial Evolution of Peptides


This technology is an in silico design tool for developing antimicrobial peptides with applications in drug development.

Problem Addressed

Antibiotic resistant bacteria are an urgent threat to modern medicine. Worldwide there are 30 million new antibiotic-resistant infections diagnosed every year and 5 million of these cases will result in death. Furthermore, the emergence of antibiotic resistant bacteria currently outpaces new drug development, and there is therefore a desperate need to accelerate the rate of antibiotic drug development. Antimicrobial peptides (AMPs) have been proposed as a promising new antibiotic modality. AMPs are small naturally occurring proteins with potent antimicrobial properties; however, most AMPs either display only limited antimicrobial activity or are highly toxic to human cells. The high design cost to develop improved AMPs has thus far limited their use as a therapeutic agent. This technology uses computer-aided design to develop novel AMPs with improved stability, potency, and toxicity profiles.


This technology uses a computer-aided evolutionary genetic algorithm driven by a fitness function to generate synthetic peptides with optimized antimicrobial function. The program uses an AMP sequence that exists in nature to generate new variants by performing virtual crossing over and substitution mutations, and then the program selects the ‘fittest’ AMP variants. Several iterations of these steps are performed then the top solutions can be synthesized for in vitro and in vivo testing. As proof of principle, the inventors used this in silico design tool to generate artificial AMPs, called guavanins, derived from the guava peptide Pg-AMP1. The guavanins designed using this computational tool displayed improved antibiotic potency against the Gram-negative bacteria P. aeruginosa, E. coli, and A. baumannii, and were additionally less toxic to human cells than the parent compound. In a mouse skin infection model, the lead guavanin compound led to a thousand-fold reduction in P. aeruginosa compared to the parent molecule. In conclusion, this computer-aided design platform is a promising new tool for generating novel AMPs with improved therapeutic properties.


  • Reduces time and cost associated with generating novel antimicrobial peptides
  • Generates antimicrobial peptides with increased antimicrobial potency and safer toxicity profiles