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

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

doi:10.3389/fmicb.2021.725727

Cited by 45 publications

(63 citation statements)

References 41 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The problem of modelling AMPs was undertaken by a number of different computational approaches. One group of these approaches are classifiers, which take a peptide as input and their task is to predict whether the peptide is an AMP or not [6, 7, 8, 9], whether it is toxic [10, 11], or whether it is active [12, 13]. A related group of methods are quantitative structure–activity relationship (QSAR) models [14, 15], which identify a set of structural features for a given peptide, all of which are associated with the peptides being AMP, for example helical structure, amphipathicity, etc.…”

Section: Introductionmentioning

confidence: 99%

Discovering highly potent antimicrobial peptides with deep generative model HydrAMP

Szymczak

Możejko

Grzegorzek

et al. 2022

Preprint

View full text Add to dashboard Cite

Antimicrobial peptides (AMP) emerge as compounds that can alleviate the global health hazard of antimicrobial resistance. Since the repertoire of experimentally verified AMPs is limited, there is a need for novel computational approaches to peptide generation. For such approaches, exploring the amino-acid peptide representation space is infeasible due to its sparsity and combinatorial complexity. Thus, we propose HydrAMP, a conditional variational autoencoder that learns a lower-dimensional and continuous space of peptides' representations and captures their antimicrobial properties. HydrAMP outperforms other approaches in generating peptides, either de novo, or by analogue discovery, and leverages parameter-controlled creativity. The model disentangles the latent representation of a peptide from its antimicrobial conditions, allowing for targeted generation. Wet-lab experiments and molecular dynamics simulation confirm the increased activity of a Pexiganan-based analogue produced by HydrAMP. HydrAMP proposes new promising AMP candidates, enabling progress towards a new generation of antibiotics.

show abstract

Section: Introductionmentioning

confidence: 99%

Discovering highly potent antimicrobial peptides with deep generative model HydrAMP

Szymczak

Możejko

Grzegorzek

et al. 2022

Preprint

View full text Add to dashboard Cite

show abstract

“…For this reason, rAMPage employs AMPlify as its AMP prediction step, and will continue to until it is surpassed in performance. Machine learning in AMP discovery is a dynamic study, ranging from AMP sequence prediction and structure classification to de novo AMP sequence generation and design [ 45 , 46 , 47 ]. While there are existing methods to mine protein databases [ 48 , 49 ], rAMPage is an all-in-one tool to mine next-generation sequencing data directly from reads to AMP prediction.…”

Section: Discussionmentioning

confidence: 99%

Mining Amphibian and Insect Transcriptomes for Antimicrobial Peptide Sequences with rAMPage

et al. 2022

View full text Add to dashboard Cite

Antibiotic resistance is a global health crisis increasing in prevalence every day. To combat this crisis, alternative antimicrobial therapeutics are urgently needed. Antimicrobial peptides (AMPs), a family of short defense proteins, are produced naturally by all organisms and hold great potential as effective alternatives to small molecule antibiotics. Here, we present rAMPage, a scalable bioinformatics discovery platform for identifying AMP sequences from RNA sequencing (RNA-seq) datasets. In our study, we demonstrate the utility and scalability of rAMPage, running it on 84 publicly available RNA-seq datasets from 75 amphibian and insect species—species known to have rich AMP repertoires. Across these datasets, we identified 1137 putative AMPs, 1024 of which were deemed novel by a homology search in cataloged AMPs in public databases. We selected 21 peptide sequences from this set for antimicrobial susceptibility testing against Escherichia coli and Staphylococcus aureus and observed that seven of them have high antimicrobial activity. Our study illustrates how in silico methods such as rAMPage can enable the fast and efficient discovery of novel antimicrobial peptides as an effective first step in the strenuous process of antimicrobial drug development.

show abstract

“…We used VAEs because they have previously been used for de novo AMP design 16,17,19,44 . The generative VAE consists of an encoder, a latent vector, and a decoder.…”

Section: Generator Variational Autoencoder (Vae)mentioning

confidence: 99%

Cell-free biosynthesis combined with deep learning accelerates de novo-development of antimicrobial peptides

Pandi

Adam

Zare

et al. 2022

Preprint

View full text Add to dashboard Cite

Bioactive peptides are key molecules in health and medicine. Deep learning holds a big promise for the discovery and design of bioactive peptides. Yet, suitable experimental approaches are required to validate candidates in high throughput and at low cost. Here, we established a cell-free protein synthesis (CFPS) pipeline for the rapid and inexpensive production of antimicrobial peptides (AMPs) directly from DNA templates. To validate our platform, we used deep learning to design thousands of AMPs de novo. Using computational methods, we prioritized 500 candidates that we produced and screened with our CFPS pipeline. We identified 30 functional AMPs, which we characterized further through molecular dynamics simulations, antimicrobial activity and toxicity. Notably, six de novo-AMPs feature broad-spectrum activity against multidrug-resistant pathogens and do not develop bacterial resistance. Our work demonstrates the potential of CFPS for production and testing of bioactive peptides within less than 24 hours and <10$ per screen.

show abstract

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

Cited by 45 publications

References 41 publications

Discovering highly potent antimicrobial peptides with deep generative model HydrAMP

Discovering highly potent antimicrobial peptides with deep generative model HydrAMP

Mining Amphibian and Insect Transcriptomes for Antimicrobial Peptide Sequences with rAMPage

Cell-free biosynthesis combined with deep learning accelerates de novo-development of antimicrobial peptides

Contact Info

Product

Resources

About