Interrogation of the integrated mobile genetic elements in gut-associated Bacteroidaceae with a consensus prediction approach

Exploration of mobile genetic element (MGE) diversity and relatedness is vital to understanding microbial communities, especially the gut microbiome, where the mobilization of antibiotic resistance and pathogenicity genes has important clinical consequences. Current MGE prediction tools are biased toward elements similar to previously-identified MGEs, especially tailed phages of proteobacterial hosts. Further, there is a need for methods to examine relatedness and gene sharing among MGEs. We present VICSIN, a consensus approach for MGE prediction and clustering of predictions to provide classification. Testing of VICSIN on datasets of Pseudomonas aeruginosa and Bacteroides fragilis genomes suggests VICSIN is the optimal approach to predict integrated MGEs from poorly-explored host taxa, because of its increased sensitivity and accuracy. We applied VICSIN to a dataset of gut-associated Bacteroidaceae genomes, identifying 816 integrated MGEs falling into 95 clusters, most of which are novel. VICSINs fast and simple network-building scheme revealed a high degree of gene sharing within and between related MGE clusters. Shared gene functions across MGEs include core mobilization functions and accessory gene content, such as type VI secretion systems and antibiotic resistance genes. The MGEs identified here encode a large portion of unknown gene content, emphasizing the fact that the full diversity of MGEs and the factors they encode remain very poorly understood. Together, this work motivates more exploration of the gut mobilome, which is likely one of the most potent drivers of microbial evolution in the human microbiome. IMPORTANCEMobile genetic elements (MGEs), including phages and integrative and conjugative elements (ICEs), drive the diversity and function of microbial communities through horizontal gene transfer. Current tools to predict MGEs in genomic sequence data are highly focused on phages, and are biased against the discovery of novel MGEs. We present VICSIN, a consensus approach to MGE prediction that is able to find a diversity of MGEs, particularly in poorly-understood bacterial taxa. By applying VICSIN to a large database of diverse Bacteroidaceae genomes, we have been able to get a distinct view of the gut mobilome, extending beyond the phageome. These novel MGEs belong to related groups, sharing a significant amount of functional gene content within and between groups, supporting a mosaic model of evolution for ICEs. Understanding how phages evolve in Bacteroidaceae hosts, however, remains elusive and highlights the need for more experimental research.