Genome Informatics and Machine Learning-Based Identification of Antimicrobial Resistance-Encoding Features and Virulence Attributes in Escherichia coli Genomes Representing Globally Prevalent Lineages, Including High-Risk Clonal Complexes
Sabiha Shaik, Anuradha Singh, Arya Suresh, Niyaz Ahmed
Abstract
With the leap in whole-genome data being generated, the application of relevant methods to mine biologically significant information from microbial genomes is of utmost importance to public health genomics. Machine-learning methods have been used not only to mine, curate, or classify the data but also to identify the relevant features that could be linked to a particular class/target. This is perhaps one of the pioneering studies that has attempted to classify a large repertoire of E. coli genome data sets (5,653 genomes) belonging to 19 different STs (including well-studied as well as understudied STs) using machine learning approaches. Important features identified by these approaches have revealed ST-specific signature proteins, which could be further studied to predict possible associations with the phenotypic profiles, thereby providing a better understanding of virulence and the resistance mechanisms among different clonal lineages of E. coli.