High Dimensional Immune Profiling Reveals CD39 as a Correlate of Tuberculosis Disease Severity - Data and Analysis
Description
This dataset contains the raw data, processed data, and code required the replicate Watt et al., 2026 (doi: https://doi.org/10.64898/2026.07.01.735885). Abstract: Immune biomarkers of tuberculosis (TB) disease severity present a challenging area of research that remains poorly understood. New technologies are able to perform larger, unbiased studies that can unravel the complex host-pathogen dynamics occurring during a Mycobacterium tuberculosis infection, the causative agent of TB. In this study, we designed a high dimensional approach combining viable bacterial burden (CFU, colony forming units) with time-of-flight mass cytometry (CyTOF) analysis to profile differences in cell-type abundance and cell-type specific protein expression during states of low, intermediate and high TB disease burden. Broadly, we segregated cell-type specific immune responses into those driven by bacterial burden and/or the mycobacterial infection strain. Interrogating these immune signatures allowed us to identify ATP-catabolizing protein CD39 as a correlate of disease severity. Treatment of mice with a small molecule inhibitor of CD39 promoted effector T cell functions and CD4 T cell expansion during Mtb infection. Collectively, our data defines the differential lung immune environment between various mycobacterial disease severity states and uncovers a potential immune biomarker of infection and therapeutic immunomodulating target to aid in the treatment of TB.
Files
Steps to reproduce
Watt et al. 2026: High Dimensional Immune Profiling Reveals CD39 as a Correlate of Tuberculosis Disease Severity - Data and Analysis This data package contains the raw and processed data for the CyTOF analysis in this study. This dataset is split into two portions: the non-perfused lung experiment, which is the primary experiment explored in this study, and the perfused lung experiment, which is our replicate experiment that excludes circulating immune cells by perfusing the lung with PBS. These analyses are split into three scripts. The first is 1_{perfusion_status} _catalyst_debarcode.R, which uses the “CATALYST” R package to prep, normalize, and debarcode the raw FCS output of the CYTOF experiment. This script follows the CATALYST debarcode tutorials. The second is 2_{perfusion_status}_debarcoded_to_processed_sce_object.R, which follows the cytofWorkflow R tutorial, using the flowCore and CATALYST R packages to convert the debarcoded fcs files into a fully normalized sce object with metadata, to allow for visualization, pseudobulking, and differential analysis. This script also follows the workflow, with the same normalization procedures, and pseudobulking and differential analysis with diffcyt. We added a differential cell-type proportion analysis using the ANOVA to test the differences in cell-type proportions in three groups. The third and final script for each dataset is 3_{perfusion_status} _celltype_and_marker_dynamics.R, this script uses the final normalized sce object to generate the plots and analyses included in this study. These include the UMAPs investigating specific cell-types and activity markers, the comparison between the perfused and non-perfuse datasets, and the mapping of differential analysis to biological interpretation (Mtb effect, virulence effect, Mtb+virulence etc.) In addition to these scripts, we include relevant data, also separated into perfused and non_perfused data 1. Each data/ directory includes {perfused_status}_lung_final.rds, which contains the final normalized sce object for each dataset set up to interact with the functions in CATALYST and flowCore, as well as in different single-cell analyses such as scMappR. 2. The data/raw directory includes the direct non-debarcoded fcs file from the helios2 3. The data/debarcoded directory includes the output of 1_{perfusion_status} _catalyst_debarcode.R 4. The data/processed directory includes the output of 2_{perfusion_status}_debarcoded_to_processed_sce_object.R, which includes all of the standard QC outputs and pseudobulked counts in “cytofWorkflow”. 5. The data/dynamics directory includes the outputs of is 3_{perfusion_status} _celltype_and_marker_dynamics.R, which shows the raw outputs of the figure panels in this study. 6. The data/keys directory includes the debarcoding template, marker panel, and cell-type marker names for each dataset.
Institutions
- University of TorontoOntario, Toronto
- Hospital for Sick ChildrenOntario, Toronto