Project Summary The rapid diagnosis of bacterial pathogens in septic patients is critical for early treatment decision making. Delayed diagnoses lead to a high rate of unnecessary antibiotic prescriptions and worse patient outcomes. One approach that has been used to improve sepsis diagnoses is circulating free DNA (cfFNA). Here, bacterial DNA present in serum is used to identify microbial pathogens and inform antibiotic treatment decisions. Unfortunately, while cfDNA approaches are good at identifying some sepsis pathogens, cfDNA does poorly at distinguishing bacterial infection from bacterial colonization. In settings where substantial background signal exists from related bacteria present in the gut, skin or lungs, existing approaches often confuse colonization as infection. Similarly, sensitivity can also be compromised by misinterpreting infection as colonization. We propose that this lack of resolution exists because cfDNA diagnostics only identify bacteria at the species level. They are unable to provide insight into the bacterial strain dynamics that underly infection. To address this, we have identified a novel approach for improving the performance of cfDNA in sepsis using bacteriophage –viruses produced by bacteria. Because bacteriophages are exquisitely specific to their particular host strain, the quantification of unique phages can provide insights into bacterial population dynamics at the strain level. Our preliminary data reveal that bacteriophage sequences are present in cfDNA collected from individuals with sepsis. Using previously collected and sequenced cfDNA data, we find that we can diagnose Pseudomonas aeruginosa infections using unique phage sequences that were not possible to diagnose with bacterial sequences alone. It may be possible to extend this approach to work with other bacterial pathogens. However, first we must develop robust computational pipelines for studying phages in cfDNA as the existing algorithms are designed to work with human and bacterial DNA. A further issue with cfDNA sequencing is speed. For cfDNA to be helpful in identifying sepsis pathogens, delivering timely results is critical. To this end, newer nanopore technologies have advantages over more time consuming illumina sequencing methods. However, the utility of nanopore sequencing for phage cfDNA is untested. We must demonstrate that nanopore sequencing of phages is both accurate and timely. Our hypothesis is that bacteriophage cfDNA can provide insight into the bacterial pathogenesis of sepsis. To test this, in the R21 portion of this grant we will develop computational protocols for studying phages in existing sepsis cfDNA datasets. Then, in the R33 portion of these studies we will develop rapid sequencing protocols for characterizing phage cfDNA in existing sepsis biorepository samples Together, these studies will generate the tools and conceptual frameworks needed to investigate the role of bacterial strains in sepsis. Moreover, these ...