Machine Learning and Network Analysis to Predict Hypothetical Protein Functions of Aeromonas hydrophila

Abstract

Aeromonas hydrophila, antibiotic resistant gram negative bacteria, is a major fish pathogen. Moreover, A. hydrophila is considered to cause 13% of gastroenteritis cases in the United States. Therefore, it is important to identify groups of proteins that are effective in antibiotic resistance and causing mortality in aquaculture. We train machine learning models on existing A. hydrophila genomes to predict functions of 83 carefully filtered hypothetical proteins. Network analysis is conducted to cluster these proteins based on their similarities. Both ML and network analysis inform about possible roles of these proteins in vaccine candidacy and fish mortality.