Person:

Bhattacharyya, Roby

Low-cost shotgun DNA sequencing is transforming the microbial sciences. Sequencing instruments are so effective that sample preparation is now the key limiting factor. Here, we introduce a microfluidic sample preparation platform that integrates the key steps in cells to sequence library sample preparation for up to 96 samples and reduces DNA input requirements 100-fold while maintaining or improving data quality. The general-purpose microarchitecture we demonstrate supports workflows with arbitrary numbers of reaction and clean-up or capture steps. By reducing the sample quantity requirements, we enabled low-input (∼10,000 cells) whole-genome shotgun (WGS) sequencing of Mycobacterium tuberculosis and soil micro-colonies with superior results. We also leveraged the enhanced throughput to sequence ∼400 clinical Pseudomonas aeruginosa libraries and demonstrate excellent single-nucleotide polymorphism detection performance that explained phenotypically observed antibiotic resistance. Fully-integrated lab-on-chip sample preparation overcomes technical barriers to enable broader deployment of genomics across many basic research and translational applications.

Dysregulation of the immune response to bacterial infection can lead to sepsis, a condition with high mortality. Multiple whole-blood gene expression studies have defined sepsis-associated molecular signatures but did not resolve changes in transcriptional states of specific cell types. Here, we used single-cell RNA sequencing to profile the blood of patients with sepsis (n = 29) across three clinical cohorts with corresponding controls (n = 36). We profiled total peripheral blood mononuclear cells (PBMCs, 106,545 cells) and dendritic cells (19,806 cells) across all patients and, based on clustering of their gene expression profiles, defined 16 immune cell states. We identified a unique CD14+ monocyte state that is expanded in septic patients and validated its power in discriminating septic patients from controls using public transcriptomic data from patients of different disease etiologies and multiple geographic locations (18 cohorts, n = 1,213 patients). We identified a panel of surface markers for isolation and quantification of the monocyte state, characterized its epigenomic and functional phenotypes, and propose a model for its induction from human bone marrow. This study demonstrates the utility of single cell genomics in discovering disease-associated cytologic signatures and provides insight into the cellular basis of immune dysregulation in bacterial sepsis.