Abstract

Methods and systems for designing phage therapy cocktails. A candidate phage selector module may be configured to predict lytic activity of phages using a first trained machine learning algorithm, and select candidate phages based at least in part on predicted lytic activity. A cocktail generator module may be configured to generate a set of phage cocktails using a second trained machine learning algorithm, based at least in part on predicted lytic activity. A feedback panel module may be configured to receive feedback information associated with phage cocktails, and use the feedback information to train the first and/or second trained machine learning algorithms.