The final, formatted version of the article will be published soon.
ORIGINAL RESEARCH article
Front. Microbiol.
Sec. Phage Biology
Volume 15 - 2024 |
doi: 10.3389/fmicb.2024.1446097
This article is part of the Research Topic Engineering Antibodies and Bacteriophages to Manage and Diagnose Microbial Infections View all 7 articles
PhageScanner: a Reconfigurable Machine Learning Framework for Bacteriophage Genomic and Metagenomic Feature Annotation
Provisionally accepted- University of Colorado Boulder, Boulder, United States
Bacteriophages are the most prolific organisms on Earth, yet many of their genomes and assemblies from metagenomic sources lack protein sequences with identified functions. While most bacteriophage proteins are structural proteins, categorized as Phage Virion Proteins (PVPs), a considerable number remain unclassified. Complicating matters further, traditional lab-based methods for PVP identification can be tedious. To expedite the process of identifying PVPs, machine-learning models are increasingly being employed. Existing tools have developed models for predicting PVPs from protein sequences as input. However, none of these efforts have built software allowing for both genomic and metagenomic data as input. In addition, there is currently no framework available for easily curating data and creating new types of machine learning models.In response, we introduce PhageScanner, an open-source platform that streamlines data collection for genomic and metagenomic datasets, model training and testing, and includes a prediction pipeline for annotating genomic and metagenomic data.PhageScanner also features a graphical user interface (GUI) for visualizing annotations on genomic and metagenomic data. We further introduce a BLAST-based classifier that outperforms ML-based models and an efficient Long Short-Term Memory (LSTM) classifier. We then showcase the capabilities of PhageScanner by predicting PVPs in six previously uncharacterized bacteriophage genomes. In addition, we create a new model that predicts phage-encoded toxins within bacteriophage genomes, thus displaying the utility of the framework.
Keywords: Bacteriophages, machine learning, phage virion proteins (PVP), Protein prediction, deep learning, PVP identification
Received: 09 Jun 2024; Accepted: 23 Aug 2024.
Copyright: © 2024 Albin, Ramsahoye, Kochavi and Alistar. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
* Correspondence:
Dreycey Albin, University of Colorado Boulder, Boulder, United States
Michelle Ramsahoye, University of Colorado Boulder, Boulder, United States
Eitan Kochavi, University of Colorado Boulder, Boulder, United States
Mirela Alistar, University of Colorado Boulder, Boulder, United States
Disclaimer: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.