PepVAE: Variational Autoencoder Framework for Antimicrobial Peptide Generation and Activity Prediction.

Scott N Dean,Jerome Anthony E Alvarez,Dan Zabetakis,Scott A Walper,Anthony P Malanoski

doi:10.3389/fmicb.2021.725727

Scott N Dean, Jerome Anthony E Alvarez + Show 3 more

Open Access

https://doi.org/10.3389/fmicb.2021.725727

Copy DOI

Abstract

New methods for antimicrobial design are critical for combating pathogenic bacteria in the post-antibiotic era. Fortunately, competition within complex communities has led to the natural evolution of antimicrobial peptide (AMP) sequences that have promising bactericidal properties. Unfortunately, the identification, characterization, and production of AMPs can prove complex and time consuming. Here, we report a peptide generation framework, PepVAE, based around variational autoencoder (VAE) and antimicrobial activity prediction models for designing novel AMPs using only sequences and experimental minimum inhibitory concentration (MIC) data as input. Sampling from distinct regions of the learned latent space allows for controllable generation of new AMP sequences with minimal input parameters. Extensive analysis of the PepVAE-generated sequences paired with antimicrobial activity prediction models supports this modular design framework as a promising system for development of novel AMPs, demonstrating controlled production of AMPs with experimental validation of predicted antimicrobial activity.

Highlights

Many pathogenic bacteria are resistant to the majority of, if not all, antibiotics that are currently being isolated using traditional methods
While the variational autoencoder (VAE) was trained on the E. coli dataset, and sampling was performed with activity against E. coli in mind, we examined the effectiveness of the generated antimicrobial peptides (AMPs) on S. aureus and P. aeruginosa, the two most common bacteria in the modified Giant Repository of AMP Activity (GRAMPA) dataset after E. coli (Supplementary Figure S1A)
This study reports the use of a peptide generation framework, PepVAE, for discovery and design of new AMP sequences

Summary

Introduction

Many pathogenic bacteria are resistant to the majority of, if not all, antibiotics that are currently being isolated using traditional methods. Unlike many small molecule antibiotics, antimicrobial peptides (AMPs), essential components of the innate immune system of humans and other organisms, have retained effectiveness as antimicrobials despite their ancient origins and widespread and continual contact with pathogens (Lazzaro et al, 2020). For this reason, among others, peptide antibiotics have been regularly deemed “drugs of last resort” for their ability to kill multidrug resistant bacteria, an increasingly important classification due to resistance formation toward conventional antibiotics (Lewies et al, 2019). Acting through mechanisms associated with membrane disruption, as well as other routes of incapacitation (Chung et al, 2015), the relative immutability of bacterial membranes and other essential AMP targets make the development of resistance to AMPs rare, but possible (Kubicek-Sutherland et al, 2016), increasing the importance of their reliable, continued discovery, to grow to the antimicrobial stockpile (Lazzaro et al, 2020)

Methods

Results

Conclusion