This repository is to replicate analyses from the manuscript titled "Integrated plasma proteomic and single-cell immune signaling network signatures demarcate mild, moderate, and severe COVID-19".
R software is needed to run the code.
Data are available from the lead author upon reasonable request.
The R pipeline “20201204 Multiomic Factor analysis.Rmd” contains code to reproduce the figures and analyses below, as well as several ancillary analyses to evaluate the quality of the multi-omic factor model. See also "20201204 Multiomic Factor analysis.html" for the report rendered from the .rmd file.
Figure 1E: Top factors discriminating Mild, Moderate, and Severe COVID-19
Figure S3: Multi-omic factor analysis (MOFA) identifies shared sources of variability across single-cell cytometry and proteomic data that are associated with COVID-19 severity
Figure S5: Correlation analysis of patient characteristics with severity
See also the R pipeline "20201204 COVID19 GAM factor smooth interaction.Rmd" for the longitudinal analysis of cytof and proteomics variables using general additive models with factor-smooth interactions.
Figure S12: Longitudinal representation of plasma proteomic and single cell immune signatures over the course of COVID-19 disease.
R pipelines for the predictive modelling of COVID-19 severity are available here: https://github.com/julienhed/COVID-Severity