paperKB
coga / coga-kb
Help
Sign in

Chunk #19 — Results — scCODA detects cell-type changes in COVID-19 patients that were not detected with non-compositional tests but confirmed in larger-scale studies

Source
scCODA is a Bayesian model for compositional single-cell data analysis.
Embedded
yes

Text

Next, we reanalyzed a recent COVID-19 single-cell study comparing compositional changes of major cell types in bronchoalveolar lavage fluid between healthy controls (n = 4), severe (n = 6) and moderate (n = 3) COVID-19 cases4 using plasma as manually selected reference (Fig. 3f and Supplementary Data 4). The study originally reported significant differential changes in pDC’s in healthy vs moderate and moderate vs severe, respectively, depletion in mDCs in severe vs healthy, and depletion of T cells in severe cases vs. moderate cases using a t test without multiple testing correction. Correcting for multiple testing resulted in only pDC’s reported as significantly changing in healthy vs mild and mild vs severe cases, respectively. scCODA confirmed the differential change in T cells, and identified a credible increase in NK cells between mild vs healthy cases, credible depletions of T cells between moderate vs severe cases, as well as a credible increase of neutrophils in healthy and moderate vs severe at an FDR level of 0.2 using Plasma as reference. For comparison, ANCOM identified significant changes in mDCs between healthy and