Fig. 1

Comparisons of the proteomic profiles of serum EVs by COVID-19 severity and identification of MACROH2A1. a Analytical procedures to identify COVID-19 refractory biomarkers. b A volcano plot of all identified 2376 serum EV proteins with ≥ 2 total unique peptide count by a non-targeted proteomic analysis comparing Group 2 and 3 cases of COVID-19. c Principal component analysis on 12 biomarker candidate proteins distinguishing Group 3 from Group 2. A scatter plot indicates Group 2–Group 3 classes based on the first 2 principal components of the 12 proteins. d Top 10 canonical pathways after Ingenuity Pathway Analysis (IPA) for EV proteins significantly upregulated with > 1.5-fold changes or downregulated with < 0.67-fold changes in comparison of Group 2 and 3 cases of COVID-19. e Top 10 disease and function after IPA for the EV proteins same as d. f A volcano plot of all identified 2376 serum EV proteins with ≥ 2 total unique peptide count by a non-targeted proteomic analysis comparing Groups 2 and 3 with Group 1 cases of COVID-19. g Principal component analysis performed on 174 biomarker candidate proteins distinguishing Groups 2 and 3 from Group 1. A scatter plot indicates Groups 2 and 3–Group 1 classes based on the first 2 principal components of the 174 proteins. h Top 10 canonical pathways after IPA for the EV proteins which were significantly upregulated with > twofold changes or downregulated with < 0.5-fold changes in comparison of Group 2 and 3 cases and Group 1 cases. i Top 10 disease and function after IPA for the EV proteins same as h. j Linear regression analysis adjusted for age and sex to examine whether the proteins with fold change > 2 or < 0.5 and p < 0.05 in both b and f significantly increased or decreased in the order of healthy controls, Group 1, and Groups 2 and 3. β: regression coefficient, FDR: false discovery rate