Abstract
Bloodstream infection is the most common infectious complication in hematopoietic cell transplantation recipients. To evaluate the genomic concordance of bloodstream pathogens and bacterial strains within the intestinal microbiome using whole genome sequencing, we developed StrainSifter, a bioinformatic pipeline to compare nucleotide variation between bacterial isolate strains and stool metagenomes. We applied StrainSifter to bloodstream isolates and stool metagenome samples from hematopoietic stem cell transplant recipients with bloodstream infections. StrainSifter is designed to identify single nucleotide variants between isolate and metagenomic short reads using stringent alignment, coverage, and variant frequency criteria for strain comparison. We identified enteric BSI isolates that were highly concordant with those in the gut microbiota, as well as highly concordant strains of typically non-enteric bacteria. These findings demonstrate the utility of StrainSifter in strain matching and provide a more precise investigation of the intestine as a reservoir of diverse pathogens capable of causing bloodstream infections.