Prediction of bacterial infection outcome using single cell RNA-seq analysis of human immune cells [sorted population Bulk RNA-seq]

During host-pathogen encounters, the complex interactions between different immune cell-types can determine the outcome of infection. Advances in single cell RNA-seq (scRNA-seq) allow to probe this complexity of immunity, and afforded the basis for deconvolution algorithms that infer cell-type compositions from bulk RNA-seq measurements. However, immune activation, an important aspect of immune surveillance, is not represented in current algorithms. Here, using scRNA-seq of human peripheral blood cells infected with Salmonella, we developed a novel deconvolution algorithm to infer dynamic immune states from bulk measurements. We applied our dynamic deconvolution algorithm both to cohorts of healthy individuals challenged ex vivo with Salmonella and to cohorts of tuberculosis patients during different stages of disease. We revealed cell-type specific immune responses associated not only with ex vivo infection phenotype but also with clinical disease stage. We propose that our approach provides a predictive power to identify risk for disease, and can be applied to comprehensively study human infection outcome. Overall design: PBMCs were isolated from a healthy individual and were infected ex vivo with Salmonella enterica serovar Typhimurium or with PBS as control. Monocytes and NKT cells were sorted from naïve and infected PBMCs. RNA was extracted 4 hours post infection.

13

Bossel Ben-Moshe N, Hen-Avivi S, Levitin N, Yehezkel D, Oosting M, Joosten LAB, Netea MG, Avraham R