Developing computational, experimental and ecological approaches to study integrated phages and their hosts

About This Grant

ABSTRACT Bacteriophages are integral components of the human microbiome, shaping microbial diversity and influencing host health, yet our understanding of phage-host dynamics remains limited, particularly for integrated phages. Recent discoveries reveal that lysogenic phages dominate the human gut microbiome over lytic phages, but current methods using short-read sequencing fail to capture phage integration sites and host attribution. This project will elucidate integrated phage-host dynamics in the human microbiome using innovative long-read sequencing approaches. The central hypothesis of this proposal is that integrated phages play critical roles in microbial community dynamics that can be detected, experimentally tested, and modeled to advance our understanding of microbiome function. Our preliminary data demonstrates that long-read sequencing dramatically improves phage genome assembly, host attribution, and integration site identification. Specifically, we have shown that geNomad most accurately detects phage boundaries when compared to ‘gold standard’ structural variation evidence of phage boundaries from Sniffles (DNA structural variant detector). We have also successfully propagated non-plaque-forming phages using stool-based cultures and discovered that p- crAssphage (Carjivirus communis) adopts a phage-plasmid lifestyle, challenging traditional models. Additionally, we have developed a quantitative modeling framework that revealed surprisingly low rates of phage induction in the microbiome. Building upon these findings, we will: (Aim 1) develop a computational framework to identify integrated prophages, their hosts, and integration sites from long-read metagenomic data, including a nextflow pipeline for assembly and annotation; (Aim 2) establish a stool-based culturomics platform to isolate non-plaque- forming phages and identify their bacterial hosts using meta-Hi-C sequencing and qPCR-based growth dynamics analysis; and (Aim 3) create computational models for viral-host population dynamics to test longstanding hypotheses about virus-host interactions in the human gut, extending our modeling to account for within-host spatial structure. Our multidisciplinary team brings together expertise in microbiome science, genomics and clinical applications (Bhatt), quantitative modeling and microbial biophysics (Huang), and evolutionary dynamics and microbial ecology (Good). We have collaborated successfully for several years and are supported by robust computational resources and experimental facilities. This project offers technological innovations through novel long-read sequencing pipelines and conceptual advances in understanding lysogenic phage ecology and lifestyle diversity. Successful completion will yield a suite of computational and experimental methods to identify integrated phages, their hosts, and viral-host interactions within the human microbiome, significantly enhancing our understanding of how phages impact human health and potentially informing future phage-based therapeutic strategies for microbiome modulation.