Accurate interpretation of within-host dissemination using barcoded bacteria.

Bacterial dissemination across tissues is a critically important process influencing infection outcomes. Monitoring within-host dissemination is challenging because conventional measures of bulk bacterial burden cannot distinguish between lineages that are shared between tissues and those that replicate locally. This limitation can be overcome using barcoded bacteria, where deep sequencing of the barcode locus and comparisons of barcodes between tissues define which lineages spread within the host. Numerous studies have used barcoded bacteria to generate high-resolution maps of dissemination. However, since multiple cells in the infectious inoculum can contain identical barcodes, inferences about dissemination can be confounded when distinct lineages from the inoculum with identical barcodes are observed in different tissues. Thus, even though the same barcodes can be observed in different tissues, dissemination between these tissues may not have occurred. Here, we aimed to develop an approach that would provide a solution to this confounding effect. We developed a simulation-based distance metric that quantifies the significance of observing shared barcodes between tissues. We validated this approach using simulated data sets spanning three orders of magnitude in barcode diversities and on three published experimental infection data sets. Our reanalysis reveals previously unappreciated patterns of