CoChat

Single-cell multi-omics analysis of the immune response in COVID-19

Open access

Published version of a preprint

The DOI matches its Crossref record; 82 of the 82 references it cites were found in a registry or an index.

Abstract

Analysis of human blood immune cells provides insights into the coordinated response to viral infections such as severe acute respiratory syndrome coronavirus 2, which causes coronavirus disease 2019 (COVID-19). We performed single-cell transcriptome, surface proteome and T and B lymphocyte antigen receptor analyses of over 780,000 peripheral blood mononuclear cells from a cross-sectional cohort of 130 patients with varying severities of COVID-19. We identified expansion of nonclassical monocytes expressing complement transcripts ( CD16 + C1QA/B/C + ) that sequester platelets and were predicted to replenish the alveolar macrophage pool in COVID-19. Early, uncommitted CD34 + hematopoietic stem/progenitor cells were primed toward megakaryopoiesis, accompanied by expanded megakaryocyte-committed progenitors and increased platelet activation. Clonally expanded CD8 + T cells and an increased ratio of CD8 + effector T cells to effector memory T cells characterized severe disease, while circulating follicular helper T cells accompanied mild disease. We observed a relative loss of IgA2 in symptomatic disease despite an overall expansion of plasmablasts and plasma cells. Our study highlights the coordinated immune response that contributes to COVID-19 pathogenesis and reveals discrete cellular components that can be targeted for therapy.

Overview

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1038/s41591-021-01329-2

Cites

82 of 82 references verified0 citation errors · 0 not checkable

Statistics

Figures

Fig. 1 Single-cell multi-omics analysis of PBMCs from individuals with COVID-19 and controls. a, Overview of the participants included and the samples and data collected. Figure was created using BioRender.com. b, UMAP visualization of all 781,123 cells after QC. Leiden clusters based on 5′ gene expression shown and colored by cell type. Lymph, lymphocyte; mono, monocyte; prolif, proliferating. c, Bar plot of the proportion of cell types shown in b, separated by condition and COVID-19 severity status. Hypothesis testing was performed using quasi-likelihood F-test comparing healthy controls to individuals with COVID-19 for linear trends across disease severity groups (healthy > asymptomatic > mild > moderate > severe > critical). Differentially abundant cell types were determined using a 10% FDR and are marked with an asterisk. d, Enrichment of interferon response of each cell state separated by severity. IFN response was calculated using a published gene list (GO:0034340) e, UMAP computed using batch-corrected mean staining intensities of 188 antibodies for 4,241 hyperspheres. Each hypersphere represents an area in the 188-dimensional space and is colored by significant (spatial FDR < 0.05) severity-associated changes in abundance of cells within that space.
Fig. 1 Single-cell multi-omics analysis of PBMCs from individuals with COVID-19 and controls. a, Overview of the participants included and the samples and data collected. Figure was created using BioRender.com. b, UMAP visualization of all 781,123 cells after QC. Leiden clusters based on 5′ gene expression shown and colored by cell type. Lymph, lymphocyte; mono, monocyte; prolif, proliferating. c, Bar plot of the proportion of cell types shown in b, separated by condition and COVID-19 severity status. Hypothesis testing was performed using quasi-likelihood F-test comparing healthy controls to individuals with COVID-19 for linear trends across disease severity groups (healthy > asymptomatic > mild > moderate > severe > critical). Differentially abundant cell types were determined using a 10% FDR and are marked with an asterisk. d, Enrichment of interferon response of each cell state separated by severity. IFN response was calculated using a published gene list (GO:0034340) e, UMAP computed using batch-corrected mean staining intensities of 188 antibodies for 4,241 hyperspheres. Each hypersphere represents an area in the 188-dimensional space and is colored by significant (spatial FDR < 0.05) severity-associated changes in abundance of cells within that space.CC BY · PubMed Central
Fig. 2 Expansion of complement-expressing nonclassical monocytes and megakaryocyte-primed progenitor cells and increased platelet activation with COVID-19 disease severity. a, Dot plots of gene (left) and surface protein (right) expression for myeloid populations. b, Bar plot of the proportion of myeloid populations from the Newcastle and London sites. Hypothesis testing was performed using a quasi-likelihood F-test comparing healthy controls to individuals with COVID-19. Differentially abundant cell types were determined using a 10% FDR and are marked with an asterisk. c, PAGA graph representing connectivity between clusters defined in a for healthy (top left) and COVID-19 (bottom left) monocytes and BAL macrophages (mac). Expression of IL6 (top right) and TNF (bottom right) in each cluster along the predicted path for COVID-19 monocytes. d, Expression of differentially expressed cytokines between CD83+CD14+ monocytes and BAL macrophages shown by cells ordered by pseudotime calculated for cells from c. e, Dot plot of gene expression of DC-derived T cell polarizing cytokines in peripheral blood DC2 cells and mature BAL DCs. f, Heat map displaying gene-set enrichment scores for type 1/3 IFN response, TNF response and JAK–STAT signatures in the myeloid populations. Asterisks indicate significance compared to healthy controls. Absolute values and other comparisons are provided in Supplementary Table 7. g, Heat map of predicted ligand–receptor interactions between platelets and monocyte subsets, using RNA data. h, Dot plot of significant differentially expressed genes between samples from healthy donors and individuals with COVID-19 filtered for platelet activation markers. i, UMAP representation of HSPCs (top) and dot plot of gene expression markers used to annotate clusters (bottom). MK, megakaryocyte; prog, progenitor. j, Bar chart depicting the proportion of progenitors. k, Box plots displaying the enrichment of a megakaryocyte progenitor signature in CD34+CD38+ HSPCs (right) and CD34+CD38− (left), averaged per donor scores. Comparisons were made by an analysis of variance (ANOVA) with pairwise comparisons using Tukey’s test. Asterisks above bars indicate significance and are colored by the severity for which they were compared to. Absolute values are provided in Supplementary Table 8. Boxes denote the interquartile range (IQR), and the median is shown as horizontal bars. Whiskers extend to 1.5 times the IQR, and outliers are shown as individual points (P values: CD38-negative cells in healthy versus LPS group (90 min), 0.3 × 10−3; CD38-positive cells in healthy versus moderate group, 0.7 × 10−3).
Fig. 2 Expansion of complement-expressing nonclassical monocytes and megakaryocyte-primed progenitor cells and increased platelet activation with COVID-19 disease severity. a, Dot plots of gene (left) and surface protein (right) expression for myeloid populations. b, Bar plot of the proportion of myeloid populations from the Newcastle and London sites. Hypothesis testing was performed using a quasi-likelihood F-test comparing healthy controls to individuals with COVID-19. Differentially abundant cell types were determined using a 10% FDR and are marked with an asterisk. c, PAGA graph representing connectivity between clusters defined in a for healthy (top left) and COVID-19 (bottom left) monocytes and BAL macrophages (mac). Expression of IL6 (top right) and TNF (bottom right) in each cluster along the predicted path for COVID-19 monocytes. d, Expression of differentially expressed cytokines between CD83+CD14+ monocytes and BAL macrophages shown by cells ordered by pseudotime calculated for cells from c. e, Dot plot of gene expression of DC-derived T cell polarizing cytokines in peripheral blood DC2 cells and mature BAL DCs. f, Heat map displaying gene-set enrichment scores for type 1/3 IFN response, TNF response and JAK–STAT signatures in the myeloid populations. Asterisks indicate significance compared to healthy controls. Absolute values and other comparisons are provided in Supplementary Table 7. g, Heat map of predicted ligand–receptor interactions between platelets and monocyte subsets, using RNA data. h, Dot plot of significant differentially expressed genes between samples from healthy donors and individuals with COVID-19 filtered for platelet activation markers. i, UMAP representation of HSPCs (top) and dot plot of gene expression markers used to annotate clusters (bottom). MK, megakaryocyte; prog, progenitor. j, Bar chart depicting the proportion of progenitors. k, Box plots displaying the enrichment of a megakaryocyte progenitor signature in CD34+CD38+ HSPCs (right) and CD34+CD38− (left), averaged per donor scores. Comparisons were made by an analysis of variance (ANOVA) with pairwise comparisons using Tukey’s test. Asterisks above bars indicate significance and are colored by the severity for which they were compared to. Absolute values are provided in Supplementary Table 8. Boxes denote the interquartile range (IQR), and the median is shown as horizontal bars. Whiskers extend to 1.5 times the IQR, and outliers are shown as individual points (P values: CD38-negative cells in healthy versus LPS group (90 min), 0.3 × 10−3; CD38-positive cells in healthy versus moderate group, 0.7 × 10−3).CC BY · PubMed Central
Fig. 3 Compositional and clonal analyses of T lymphocytes illustrate the expansion of effector subsets. a, UMAP visualization of 309,617 T cells based on gene expression shown and colored by cell type. Insets show the two-dimensional kernel density estimates of select T cell types in UMAP space. b, Dot plots of gene (top) and surface protein (bottom) expression for populations shown in a. c, Dot plots of gene expression of cytokine genes for populations shown in a. d, Box plots of cell type proportions that are differentially abundant between healthy donors and individuals with COVID-19. Boxes denote the IQR, and the median is shown as horizontal bars. Whiskers extend to 1.5 times the IQR and outliers are shown as individual points (n = 24 healthy, n = 86 COVID-19 biologically independent samples). e, Box plots of the proportion of cell types shown in a. Only cell types showing trends of changes by severity status are shown. Boxes denote IQR with median shown as horizontal bars. Whiskers extend to 1.5 times the IQR, and outliers are shown as individual points (n = 9 asymptomatic, n = 23 mild, n = 30 moderate, n = 13 severe, n = 10 critical biologically independent samples). f, Bar plots show the frequency of clonal T cells. Expanded clones denote clonotypes observed more than once. Asterisks indicate significance after multiple-testing correction (logistic regression using two-sided t-test with Benjamini–Hochberg FDR correction; CD4+ TCM adjusted P = 0.119, CD4+ TEM adjusted P = 0.472, CD4+IL-22+ adjusted P = 0.01, CD4+ prolif. adjusted P = 0.993, CD4+ TH1 adjusted P = 0.993, CD4+ TFH adjusted P = 0.109, Treg adjusted P = 0.993, CD8+ prolif. adjusted P = 0.016, CD8+ TTE adjusted P = 2.49 × 10−15, CD8+ TEM adjusted P = 0.259). g, Box plots of the proportion of clonally expanded CD8+ TEM cells (left), effector CD8+ T cells (middle) and the ratio of effector CD8+ T cells to CD8+ TEM cells (right). Boxes denote the IQR, and the median is shown as horizontal bars. Whiskers extend to 1.5 times the IQR, and outliers are shown as individual points. Legend is as in e.
Fig. 3 Compositional and clonal analyses of T lymphocytes illustrate the expansion of effector subsets. a, UMAP visualization of 309,617 T cells based on gene expression shown and colored by cell type. Insets show the two-dimensional kernel density estimates of select T cell types in UMAP space. b, Dot plots of gene (top) and surface protein (bottom) expression for populations shown in a. c, Dot plots of gene expression of cytokine genes for populations shown in a. d, Box plots of cell type proportions that are differentially abundant between healthy donors and individuals with COVID-19. Boxes denote the IQR, and the median is shown as horizontal bars. Whiskers extend to 1.5 times the IQR and outliers are shown as individual points (n = 24 healthy, n = 86 COVID-19 biologically independent samples). e, Box plots of the proportion of cell types shown in a. Only cell types showing trends of changes by severity status are shown. Boxes denote IQR with median shown as horizontal bars. Whiskers extend to 1.5 times the IQR, and outliers are shown as individual points (n = 9 asymptomatic, n = 23 mild, n = 30 moderate, n = 13 severe, n = 10 critical biologically independent samples). f, Bar plots show the frequency of clonal T cells. Expanded clones denote clonotypes observed more than once. Asterisks indicate significance after multiple-testing correction (logistic regression using two-sided t-test with Benjamini–Hochberg FDR correction; CD4+ TCM adjusted P = 0.119, CD4+ TEM adjusted P = 0.472, CD4+IL-22+ adjusted P = 0.01, CD4+ prolif. adjusted P = 0.993, CD4+ TH1 adjusted P = 0.993, CD4+ TFH adjusted P = 0.109, Treg adjusted P = 0.993, CD8+ prolif. adjusted P = 0.016, CD8+ TTE adjusted P = 2.49 × 10−15, CD8+ TEM adjusted P = 0.259). g, Box plots of the proportion of clonally expanded CD8+ TEM cells (left), effector CD8+ T cells (middle) and the ratio of effector CD8+ T cells to CD8+ TEM cells (right). Boxes denote the IQR, and the median is shown as horizontal bars. Whiskers extend to 1.5 times the IQR, and outliers are shown as individual points. Legend is as in e.CC BY · PubMed Central
Fig. 4 Single-cell analysis of B lymphocytes and BCR repertoire reveal plasmablast expansion and clonality differences between genders. a, UMAP visualization of 74,019 cells in the B cell lineage identified from gene expression data. b, Dot plots of gene (top) and surface protein (bottom) expression for populations shown in a. c, Bar plot of the mean proportion of cell types shown in a. d, Proportion of total IgA and IgA2 in plasmablast and plasma cells based on BCR data. Kruskal–Wallis test with Benjamini–Hochberg correction. e, Coordinated changes between TFH and B cells assessed by differential correlation analysis (empirical P ≤ 0.1). Shown is the Pearson correlation (± bootstrap s.e.m.) between TFH proportions and plasmablast or plasma cell (combined); only significant trends are shown. f, GSEA of MSigDB hallmark signatures in naive B cells, switched memory B cells and plasmablasts for asymptomatic/symptomatic COVID-19 versus healthy groups. Size of circles indicate (absolute) normalized enrichment score (NES). GSEA (permutation) nominal P < 0.05 and FDR < 0.25 denoted by non-gray colored dots. EMT, epithelial–mesenchymal transition; UV, ultraviolet. g, Dot plots of genes related to TNF signaling and BCR signaling in naive B cells, switched memory B cells and plasmablasts. Size of circles indicates the percentage of cells expressing the gene, and color gradient corresponds to increasing mean expression value. h, Scatterplot of clonotype size by node closeness centrality gini indices with marginal histograms indicating the distribution. Each dot represents an individual. i, BCR overlap incidence plot. Nodes correspond to individual donors colored by (inner ring) severity and (outer ring) site from which samples were collected. Edges indicate if at least one cell from each individual displayed an identical combination of heavy and light-chain V and J gene usage with CDR3 similarity allowance (≥85%). j, Clonotype size (left) and node closeness centrality gini indices (right) separated by gender. Mann–Whitney U test with Benjamini–Hochberg correction between the gender groups within each severity status. Color of adjusted P values indicates the gender group with the higher mean value. The box portion of the box plots extends from the 25th to 75th percentiles, whiskers extend from the smallest to largest values, and the middle line corresponds to the median. NS, not significant.
Fig. 4 Single-cell analysis of B lymphocytes and BCR repertoire reveal plasmablast expansion and clonality differences between genders. a, UMAP visualization of 74,019 cells in the B cell lineage identified from gene expression data. b, Dot plots of gene (top) and surface protein (bottom) expression for populations shown in a. c, Bar plot of the mean proportion of cell types shown in a. d, Proportion of total IgA and IgA2 in plasmablast and plasma cells based on BCR data. Kruskal–Wallis test with Benjamini–Hochberg correction. e, Coordinated changes between TFH and B cells assessed by differential correlation analysis (empirical P ≤ 0.1). Shown is the Pearson correlation (± bootstrap s.e.m.) between TFH proportions and plasmablast or plasma cell (combined); only significant trends are shown. f, GSEA of MSigDB hallmark signatures in naive B cells, switched memory B cells and plasmablasts for asymptomatic/symptomatic COVID-19 versus healthy groups. Size of circles indicate (absolute) normalized enrichment score (NES). GSEA (permutation) nominal P < 0.05 and FDR < 0.25 denoted by non-gray colored dots. EMT, epithelial–mesenchymal transition; UV, ultraviolet. g, Dot plots of genes related to TNF signaling and BCR signaling in naive B cells, switched memory B cells and plasmablasts. Size of circles indicates the percentage of cells expressing the gene, and color gradient corresponds to increasing mean expression value. h, Scatterplot of clonotype size by node closeness centrality gini indices with marginal histograms indicating the distribution. Each dot represents an individual. i, BCR overlap incidence plot. Nodes correspond to individual donors colored by (inner ring) severity and (outer ring) site from which samples were collected. Edges indicate if at least one cell from each individual displayed an identical combination of heavy and light-chain V and J gene usage with CDR3 similarity allowance (≥85%). j, Clonotype size (left) and node closeness centrality gini indices (right) separated by gender. Mann–Whitney U test with Benjamini–Hochberg correction between the gender groups within each severity status. Color of adjusted P values indicates the gender group with the higher mean value. The box portion of the box plots extends from the 25th to 75th percentiles, whiskers extend from the smallest to largest values, and the middle line corresponds to the median. NS, not significant.CC BY · PubMed Central
Fig. 5 Integrated framework of the peripheral immune response in COVID-19. Schematic illustration of study highlights. Created with BioRender.com.
Fig. 5 Integrated framework of the peripheral immune response in COVID-19. Schematic illustration of study highlights. Created with BioRender.com.CC BY · PubMed Central
Extended Data Fig. 1 Single-cell analysis quality control and cell type definition for COVID PBMC single cell analysis. a, Bar chart showing the composition of sample severities across the three sites (n = 23 mild, n = 30 moderate, n = 11, severe, n = 10 critical biologically independent samples). b, Scatter plot displaying the total number of gene counts per sample from each site. c, UMAP visualisations from Fig. 1b coloured by site. d, Boxplot of kBET results calculated both before and after batch correction with Harmony for each cluster in Fig. 1b kBET statistic calculation using patient ID as the batch factor (n = 130 biologically independent samples, n = 627,172 cells in 1 experiment). Hinges indicate to 25th and 75th percentile and whiskers to lowest and highest value in 1.5*interquartile range. e, Dot plots of 5’ gene expression (top; blue) and surface protein (bottom; red) expression for populations shown in Fig. 1a where the colour is scaled by mean expression and the dot size is proportional to the percent of the population expressing the gene/protein, respectively. f, Tile plot showing percentage concordance between COVID-19 PBMC annotation (x-axis) and Azimuth annotation (y-axis) (https://satijalab.org/azimuth/).
Extended Data Fig. 1 Single-cell analysis quality control and cell type definition for COVID PBMC single cell analysis. a, Bar chart showing the composition of sample severities across the three sites (n = 23 mild, n = 30 moderate, n = 11, severe, n = 10 critical biologically independent samples). b, Scatter plot displaying the total number of gene counts per sample from each site. c, UMAP visualisations from Fig. 1b coloured by site. d, Boxplot of kBET results calculated both before and after batch correction with Harmony for each cluster in Fig. 1b kBET statistic calculation using patient ID as the batch factor (n = 130 biologically independent samples, n = 627,172 cells in 1 experiment). Hinges indicate to 25th and 75th percentile and whiskers to lowest and highest value in 1.5*interquartile range. e, Dot plots of 5’ gene expression (top; blue) and surface protein (bottom; red) expression for populations shown in Fig. 1a where the colour is scaled by mean expression and the dot size is proportional to the percent of the population expressing the gene/protein, respectively. f, Tile plot showing percentage concordance between COVID-19 PBMC annotation (x-axis) and Azimuth annotation (y-axis) (https://satijalab.org/azimuth/).CC BY · PubMed Central
Extended Data Fig. 2 Differential abundance analysis and expression of GWAS hits related to cytokines, chemokines and growth factors. a, Volcano plots showing differential abundance testing performed using quasi-likelihood F-test comparing healthy controls to cases for linear trends across disease severity groups. Differentially abundant cell types were determined using a 10% false discovery rate (FDR). b, Scatter plots showing blood counts for Newcastle data grouped by severity status (n = 6 healthy and LPS samples; n = 11 mild, n = 17 moderate, n = 7 severe, and n = 7 critical biologically independent samples). Dotted lines mark the normal ranges. Significance determined using Kruskal-Wallis with Dunn’s post hoc corrected for multiple comparison. WBC, whole blood count. c, Forest plot showing the standard deviation of each clinical/technical factor estimated by the Poisson generalised linear mixed model. Error bars show the standard error estimated from the Fisher information matrix (n = 130 biologically independent samples) SD, standard deviation. d, Volcano plots showing differential abundance testing according to time since symptom onset. Differentially abundant (FDR 10%) points are shown in red. e, Box plots displaying the duration of COVID-19 symptoms from the onset by severity (n = 23 mild, n = 30 moderate, n = 11, severe, n = 10 critical biologically independent samples). Boxes denote IQR with median shown as horizontal bars. Whiskers extend to 1.5x the IQR; outliers are shown as individual points. f, Box plots displaying the duration of COVID-19 symptoms from the onset split by severity and sex (n = 23 mild, n = 30 moderate, n = 11, severe, n = 10 critical biologically independent samples). Boxes denote IQR with median shown as horizontal bars. Whiskers extend to 1.5x the IQR; outliers are shown as individual points. g, Correlated log fold-changes of cell type abundance changes as a function of symptom duration with (x-axis) and without critically ill patients (y-axis). h, Differential abundance testing with a leave-one-out analysis for the T cells (top) of B cells (bottom) (FDR10%). i, Heat map displaying fold change over healthy (left) and dot plot of gene expression where the colour is scaled by mean expression and the dot size is proportional to the percent of the population expressing the gene (right) for genes associated with COVID-19 identified in recent GWAS studies18,19 for the cell populations in Fig. 1b. j, Heat map displaying normalised values of cytokine, chemokine and growth factors in serum of patients with COVID-19.
Extended Data Fig. 2 Differential abundance analysis and expression of GWAS hits related to cytokines, chemokines and growth factors. a, Volcano plots showing differential abundance testing performed using quasi-likelihood F-test comparing healthy controls to cases for linear trends across disease severity groups. Differentially abundant cell types were determined using a 10% false discovery rate (FDR). b, Scatter plots showing blood counts for Newcastle data grouped by severity status (n = 6 healthy and LPS samples; n = 11 mild, n = 17 moderate, n = 7 severe, and n = 7 critical biologically independent samples). Dotted lines mark the normal ranges. Significance determined using Kruskal-Wallis with Dunn’s post hoc corrected for multiple comparison. WBC, whole blood count. c, Forest plot showing the standard deviation of each clinical/technical factor estimated by the Poisson generalised linear mixed model. Error bars show the standard error estimated from the Fisher information matrix (n = 130 biologically independent samples) SD, standard deviation. d, Volcano plots showing differential abundance testing according to time since symptom onset. Differentially abundant (FDR 10%) points are shown in red. e, Box plots displaying the duration of COVID-19 symptoms from the onset by severity (n = 23 mild, n = 30 moderate, n = 11, severe, n = 10 critical biologically independent samples). Boxes denote IQR with median shown as horizontal bars. Whiskers extend to 1.5x the IQR; outliers are shown as individual points. f, Box plots displaying the duration of COVID-19 symptoms from the onset split by severity and sex (n = 23 mild, n = 30 moderate, n = 11, severe, n = 10 critical biologically independent samples). Boxes denote IQR with median shown as horizontal bars. Whiskers extend to 1.5x the IQR; outliers are shown as individual points. g, Correlated log fold-changes of cell type abundance changes as a function of symptom duration with (x-axis) and without critically ill patients (y-axis). h, Differential abundance testing with a leave-one-out analysis for the T cells (top) of B cells (bottom) (FDR10%). i, Heat map displaying fold change over healthy (left) and dot plot of gene expression where the colour is scaled by mean expression and the dot size is proportional to the percent of the population expressing the gene (right) for genes associated with COVID-19 identified in recent GWAS studies18,19 for the cell populations in Fig. 1b. j, Heat map displaying normalised values of cytokine, chemokine and growth factors in serum of patients with COVID-19.CC BY · PubMed Central
Extended Data Fig. 3 Myeloid comparisons with bronchial alveolar lavage dataset and receptor-ligand interaction analysis between megakaryocyte, myeloid and progenitor cells. a, Volcano plot showing differential abundance testing according to time since symptom onset for the myeloid populations. Differentially abundant (FDR 10%) points are shown in red. b, Dot plots of gene expression of C1 complement components for cells in Fig. 1b where the colour is scaled by mean expression and the dot size is proportional to the percent of the population expressing the gene. c, Dot plots of gene expression of a recently published BAL dataset (accession number GSE14592622) for genes in Fig. 2a where the colour is scaled by mean expression and the dot size is proportional to the percent of the population expressing the gene. d, Heat map of CellPhoneDB predicted ligand:receptor interactions between platelets and monocyte subsets, based on the protein data. e, Dot plots of expression of protein markers used to annotate clusters in Fig. 2i. MK, Megakaryocyte f. Heatmap of differentially expressed genes between megakaryocyte, myeloid and erythroid progenitor clusters. MK, megakaryocyte; My, myeloid. g, Box plots displaying enrichment of an erythroid signature (top) and a myeloid signature (bottom) found in CD34−CD38− (left) and CD34+CD38+ HSPCs (right), separated by severity. Asterisks above bars indicate significance and are coloured by the severity for which they were compared to. Absolute values are provided in Supplementary Tables 8 and 9. (n = 120 biologically independent samples, n = 3297 cells in 1 experiment). Boxes denote IQR with median shown as horizontal bars. Whiskers extend to 1.5x the IQR; outliers are shown as individual points (n = 24 healthy, n = 86 COVID-19 biologically independent samples) (p-values: Myeloid signature in CD38 negative HSPCs, Healthy vs. Mild: 0.8 x 10-3, Healthy vs Moderate 0.02).
Extended Data Fig. 3 Myeloid comparisons with bronchial alveolar lavage dataset and receptor-ligand interaction analysis between megakaryocyte, myeloid and progenitor cells. a, Volcano plot showing differential abundance testing according to time since symptom onset for the myeloid populations. Differentially abundant (FDR 10%) points are shown in red. b, Dot plots of gene expression of C1 complement components for cells in Fig. 1b where the colour is scaled by mean expression and the dot size is proportional to the percent of the population expressing the gene. c, Dot plots of gene expression of a recently published BAL dataset (accession number GSE14592622) for genes in Fig. 2a where the colour is scaled by mean expression and the dot size is proportional to the percent of the population expressing the gene. d, Heat map of CellPhoneDB predicted ligand:receptor interactions between platelets and monocyte subsets, based on the protein data. e, Dot plots of expression of protein markers used to annotate clusters in Fig. 2i. MK, Megakaryocyte f. Heatmap of differentially expressed genes between megakaryocyte, myeloid and erythroid progenitor clusters. MK, megakaryocyte; My, myeloid. g, Box plots displaying enrichment of an erythroid signature (top) and a myeloid signature (bottom) found in CD34−CD38− (left) and CD34+CD38+ HSPCs (right), separated by severity. Asterisks above bars indicate significance and are coloured by the severity for which they were compared to. Absolute values are provided in Supplementary Tables 8 and 9. (n = 120 biologically independent samples, n = 3297 cells in 1 experiment). Boxes denote IQR with median shown as horizontal bars. Whiskers extend to 1.5x the IQR; outliers are shown as individual points (n = 24 healthy, n = 86 COVID-19 biologically independent samples) (p-values: Myeloid signature in CD38 negative HSPCs, Healthy vs. Mild: 0.8 x 10-3, Healthy vs Moderate 0.02).CC BY · PubMed Central
Extended Data Fig. 4 UMAP embedding of T lymphoid clusters showing cell type estimation and covariates. a, UMAP visualisation of 309,617 T cells separated by sources of donors. b, UMAP visualisation showing 2-dimensional kernel density estimates of each T cell type in UMAP space. c.-e. UMAP visualisation of T cells coloured by gender (c), disease severity status (d) and age (e).
Extended Data Fig. 4 UMAP embedding of T lymphoid clusters showing cell type estimation and covariates. a, UMAP visualisation of 309,617 T cells separated by sources of donors. b, UMAP visualisation showing 2-dimensional kernel density estimates of each T cell type in UMAP space. c.-e. UMAP visualisation of T cells coloured by gender (c), disease severity status (d) and age (e).CC BY · PubMed Central
Extended Data Fig. 5 Differential abundance testing, gene set enrichment analysis and clonal diversity analysis of T lymphoid compartment. a, Box plots showing the proportion of cell types shown in Fig. 3a, n = 108 biologically independent samples. Boxes denote IQR with median shown as horizontal bars. Whiskers extend to 1.5x the IQR; outliers are shown as individual points. b, Volcano plots showing results of differential abundance testing. Abundance counts were modelled either comparing healthy vs. COVID-19, or as a function of severity. Hypothesis testing was performed using quasi-likelihood F-test comparing healthy controls to cases, or for either a linear or quadratic trend across disease severity. Differentially abundant cell types were determined using a 10% false discovery rate (FDR). c, Heatmaps showing mean expression levels across T cell subsets for suppressive (FOXP3), proliferating (MKI67) and exhaustion markers (PDCD1, HAVCR2, LAG3, TIGIT, TOX). Columns denote the mean log-normalised expression within each severity category and healthy controls. d, Volcano plot showing differential abundance testing according to time since symptom onset for the T cell populations. Differentially abundant (FDR 10%) points are shown in red. e, Gene set enrichment (Methods) in each T cell type based on differential gene expression (DGE) analysis was performed across COVID-19 disease severity groups, ordered from healthy > asymptomatic > mild > moderate > severe > critical. Statistically significant DE genes were defined with FDR < 1%. Significant enrichments were defined with 10% FDR. f, Bar plots showing percent (mean + /- SEM) of CD3+CD4+ (blue) and CD3+CD8+ (green) T cells expressing CD107a (left) and CD137 (right) in response to SARS-CoV-2 S peptide stimulation (n = 12 healthy, n = 10 mild, n = 17 moderate, n = 9 severe, and n = 5 critical biologically independent samples). Significance determined using Kruskal-Wallis with Dunn’s post-hoc corrected for multiple comparison. g, Box plots showing clone size distribution for each T cell subset (n = 9 asymptomatic, n = 19 mild, n = 29 moderate, n = 12 severe, n = 10 critical biologically independent samples). Boxes denote IQR with median shown as horizontal bars. Whiskers extend to 1.5x the IQR; outliers are shown as individual points. h, Box plots slowing clonal diversity for each T cell subset (n = 22 healthy, n = 12 asymptomatic, n = 22 mild, n = 31 moderate, n = 14 severe, n = 13 critical biologically independent samples). Boxes denote IQR with median shown as horizontal bars. Whiskers extend to 1.5x the IQR; outliers are shown as individual points. Source data
Extended Data Fig. 5 Differential abundance testing, gene set enrichment analysis and clonal diversity analysis of T lymphoid compartment. a, Box plots showing the proportion of cell types shown in Fig. 3a, n = 108 biologically independent samples. Boxes denote IQR with median shown as horizontal bars. Whiskers extend to 1.5x the IQR; outliers are shown as individual points. b, Volcano plots showing results of differential abundance testing. Abundance counts were modelled either comparing healthy vs. COVID-19, or as a function of severity. Hypothesis testing was performed using quasi-likelihood F-test comparing healthy controls to cases, or for either a linear or quadratic trend across disease severity. Differentially abundant cell types were determined using a 10% false discovery rate (FDR). c, Heatmaps showing mean expression levels across T cell subsets for suppressive (FOXP3), proliferating (MKI67) and exhaustion markers (PDCD1, HAVCR2, LAG3, TIGIT, TOX). Columns denote the mean log-normalised expression within each severity category and healthy controls. d, Volcano plot showing differential abundance testing according to time since symptom onset for the T cell populations. Differentially abundant (FDR 10%) points are shown in red. e, Gene set enrichment (Methods) in each T cell type based on differential gene expression (DGE) analysis was performed across COVID-19 disease severity groups, ordered from healthy > asymptomatic > mild > moderate > severe > critical. Statistically significant DE genes were defined with FDR < 1%. Significant enrichments were defined with 10% FDR. f, Bar plots showing percent (mean + /- SEM) of CD3+CD4+ (blue) and CD3+CD8+ (green) T cells expressing CD107a (left) and CD137 (right) in response to SARS-CoV-2 S peptide stimulation (n = 12 healthy, n = 10 mild, n = 17 moderate, n = 9 severe, and n = 5 critical biologically independent samples). Significance determined using Kruskal-Wallis with Dunn’s post-hoc corrected for multiple comparison. g, Box plots showing clone size distribution for each T cell subset (n = 9 asymptomatic, n = 19 mild, n = 29 moderate, n = 12 severe, n = 10 critical biologically independent samples). Boxes denote IQR with median shown as horizontal bars. Whiskers extend to 1.5x the IQR; outliers are shown as individual points. h, Box plots slowing clonal diversity for each T cell subset (n = 22 healthy, n = 12 asymptomatic, n = 22 mild, n = 31 moderate, n = 14 severe, n = 13 critical biologically independent samples). Boxes denote IQR with median shown as horizontal bars. Whiskers extend to 1.5x the IQR; outliers are shown as individual points. Source dataCC BY · PubMed Central
Extended Data Fig. 6 Single-cell analysis of B lymphocytes proportion and gene set differences across COVID-19 severity. a, Heatmap of mean gene set enrichment scores of (top) adult peripheral blood B cell signatures32 and (bottom) bone marrow B cell signatures68. Row enrichment value is scaled from 0-1 and colour gradient corresponds to increasing mean enrichment score. b, (Top) Kruskal-Wallis test results with Benjamini-Hochberg false discovery correction for cell type proportion differences in plasmablast/plasma cells between severities. Significance is denoted by Benjamini-Hochberg corrected P < 0.05 (red text). (Bottom) Cell type abundance counts were modelled as a function of severity. Hypothesis testing was performed using quasi-likelihood F-test comparing asymptomatic to symptomatic covid, for either a linear or quadratic trend across severities. Differentially abundant cell types were determined using a 10% false discovery rate (FDR). c, (Left) Bar plot showing the mean proportion of plasmablast/plasma cells expressing IgA, IgD, IgE, IgG or IgM, based on V(D)J information. (Right) Proportion of IgM, IgA1, total IgG, IgG1, and IgG2 subclass in plasmablast and plasma cells. Kruskal-Wallis test results with Benjamini-Hochberg false discovery correction for cell type proportion differences in plasmablast/plasma cells between severities showed no significant differences. n = 10, 6, 21, 24, 11, 11 (IgM), n = 18, 9, 21, 27, 12, 9 (IgA1), n = 16, 8, 23, 27, 13, 8 (IgG (all)), n = 11, 6, 22, 25, 11, 8 (IgG1) and n = 8, 4, 16, 23, 9, 7 (IgG2) biologically independent patients/samples for healthy, asymptomatic, mild, moderate, severe and critical respectively. d, GSEA of pathways in B cell subsets for asymptomatic/symptomatic COVID versus healthy. Size of circles indicate (absolute) normalised enrichment score (NES). Pathways were considered statistically significant if GSEA (permutation) nominal P < 0.05 and FDR < 0.25 (denoted by non-grey coloured dots; boundary lines in the middle marks FDR = 0.25). EMT, Epithelial-mesenchymal transition. e, Dot plots of TNF signalling molecules, activating and inhibitory BCR signaling molecules (5’ gene expression data) in immature B cells, non-switched memory B cells, ‘exhausted’ B cells and plasma cells. Size of circles indicate percent of cells expressing the gene and colour gradient corresponds to increasing mean expression value.
Extended Data Fig. 6 Single-cell analysis of B lymphocytes proportion and gene set differences across COVID-19 severity. a, Heatmap of mean gene set enrichment scores of (top) adult peripheral blood B cell signatures32 and (bottom) bone marrow B cell signatures68. Row enrichment value is scaled from 0-1 and colour gradient corresponds to increasing mean enrichment score. b, (Top) Kruskal-Wallis test results with Benjamini-Hochberg false discovery correction for cell type proportion differences in plasmablast/plasma cells between severities. Significance is denoted by Benjamini-Hochberg corrected P < 0.05 (red text). (Bottom) Cell type abundance counts were modelled as a function of severity. Hypothesis testing was performed using quasi-likelihood F-test comparing asymptomatic to symptomatic covid, for either a linear or quadratic trend across severities. Differentially abundant cell types were determined using a 10% false discovery rate (FDR). c, (Left) Bar plot showing the mean proportion of plasmablast/plasma cells expressing IgA, IgD, IgE, IgG or IgM, based on V(D)J information. (Right) Proportion of IgM, IgA1, total IgG, IgG1, and IgG2 subclass in plasmablast and plasma cells. Kruskal-Wallis test results with Benjamini-Hochberg false discovery correction for cell type proportion differences in plasmablast/plasma cells between severities showed no significant differences. n = 10, 6, 21, 24, 11, 11 (IgM), n = 18, 9, 21, 27, 12, 9 (IgA1), n = 16, 8, 23, 27, 13, 8 (IgG (all)), n = 11, 6, 22, 25, 11, 8 (IgG1) and n = 8, 4, 16, 23, 9, 7 (IgG2) biologically independent patients/samples for healthy, asymptomatic, mild, moderate, severe and critical respectively. d, GSEA of pathways in B cell subsets for asymptomatic/symptomatic COVID versus healthy. Size of circles indicate (absolute) normalised enrichment score (NES). Pathways were considered statistically significant if GSEA (permutation) nominal P < 0.05 and FDR < 0.25 (denoted by non-grey coloured dots; boundary lines in the middle marks FDR = 0.25). EMT, Epithelial-mesenchymal transition. e, Dot plots of TNF signalling molecules, activating and inhibitory BCR signaling molecules (5’ gene expression data) in immature B cells, non-switched memory B cells, ‘exhausted’ B cells and plasma cells. Size of circles indicate percent of cells expressing the gene and colour gradient corresponds to increasing mean expression value.CC BY · PubMed Central
Extended Data Fig. 7 Single-cell BCR networks across COVID-19 severity. Single-cell BCR network plots for each severity status coloured by heavy chain isotype class (IgM, IgD, IgA, IgE, or IgG). Each circle/node corresponds to a single B cell with a corresponding set of BCR(s). Each clonotype is presented as a minimally connected graph with edge widths scaled to 1/d + 1 for edge weight d where d corresponds to the total (Levenshtein) edit distance of BCRs between two cells. Size of nodes is scaled according to increasing node closeness centrality scores that is nodes that are highly central to a clonotype network will be larger.
Extended Data Fig. 7 Single-cell BCR networks across COVID-19 severity. Single-cell BCR network plots for each severity status coloured by heavy chain isotype class (IgM, IgD, IgA, IgE, or IgG). Each circle/node corresponds to a single B cell with a corresponding set of BCR(s). Each clonotype is presented as a minimally connected graph with edge widths scaled to 1/d + 1 for edge weight d where d corresponds to the total (Levenshtein) edit distance of BCRs between two cells. Size of nodes is scaled according to increasing node closeness centrality scores that is nodes that are highly central to a clonotype network will be larger.CC BY · PubMed Central
Extended Data Fig. 8 Single-cell BCR clonotype expansion analysis across COVID-19 severity. a, UMAP visualisation of B cell lineage and coloured by clonotype size in the V(D)J data. Only expanded clonotypes are coloured (clonotype size > 2). b, Single-cell BCR network plots for each severity status coloured by assigned cell type. c, Single-cell BCR network plots for each severity status coloured by heavy chain isotype subclass (IgM, IgD, IgA1, IgA2, IgE, IgG1, IgG2, IgG3 or IgG4). Each circle/node corresponds to a single B cell with a corresponding set of BCR(s). Each clonotype is presented as a minimally connected graph with edge widths scaled to 1/d + 1 for edge weight d where d corresponds to the total (Levenshtein) edit distance of BCRs between two cells. Size of nodes is scaled according to increasing node closeness centrality scores that is nodes that are highly central to a clonotype network will be larger.
Extended Data Fig. 8 Single-cell BCR clonotype expansion analysis across COVID-19 severity. a, UMAP visualisation of B cell lineage and coloured by clonotype size in the V(D)J data. Only expanded clonotypes are coloured (clonotype size > 2). b, Single-cell BCR network plots for each severity status coloured by assigned cell type. c, Single-cell BCR network plots for each severity status coloured by heavy chain isotype subclass (IgM, IgD, IgA1, IgA2, IgE, IgG1, IgG2, IgG3 or IgG4). Each circle/node corresponds to a single B cell with a corresponding set of BCR(s). Each clonotype is presented as a minimally connected graph with edge widths scaled to 1/d + 1 for edge weight d where d corresponds to the total (Levenshtein) edit distance of BCRs between two cells. Size of nodes is scaled according to increasing node closeness centrality scores that is nodes that are highly central to a clonotype network will be larger.CC BY · PubMed Central
Extended Data Fig. 9 Single-cell BCR analysis between genders across COVID-19 severity. a, Summary plot for BCR IGHV gene usage split by gender (top: female; bottom: male). Each data point is presented as the mean gene usage proportion + /- standard error of the mean for samples within each sample group. Kruskal-Wallis test with Benjamini-Hochberg *P < 0.05 (adjusted p = 0.035) n = 12, 5, 8, 20, 6 and 6 biologically independent male patients/samples and n = 12, 4, 15, 10, 7 and 5 biologically independent female patients/samples for healthy, asymptomatic, mild, moderate, severe and critical respectively. b, (Top) Scatter plot of clonotype/cluster size by vertex size gini indices computed from contracted BCR networks (identical nodes are merged and counted). Each dot represents the gini indices of an individual coloured by severity status. Marginal histograms indicate the distribution of samples in a given severity status along the axes. (Bottom, left) Cluster/clonotype size (contracted network) gini indices separated by gender. (Bottom, right) Vertex size (contracted network) gini indices separated by gender. The boxes extend from the 25th to 75th percentiles. The whiskers go down to the smallest value and up to the largest. The line in the middle of the boxes is plotted at the median. Statistical tests were performed with non-parametric Mann-Whitney U test between the gender groups within each severity status and were considered statistically significant if Benjamini-Hochberg corrected P < 0.05 (denoted by *; n.s. denotes not significant). The Benjamini-Hochberg adjusted p-values are as follows: cluster/clonotype size (contracted network) gini indices comparisons: 3.42x10-1, 1.91x10-25, 7.71x10-272, 1.31x10-169, 3.84x10-113 and 2.98x10-1; vertex size (contracted network) gini indices comparisons: 1.12x10-57, 2.17x10-1, <1x10-308, 3.64x10-49, 2.69x10-53 and 2.37x10-04 - for healthy, asymptomatic, mild, moderate, severe and critical respectively. Colour of asterisks indicates which gender group displays a higher mean gini index (yellow: female; grey: male). n = 12, 5, 8, 20, 6 and 6 biologically independent male patients/samples and n = 12, 4, 15, 10, 7 and 5 biologically independent female patients/samples for healthy, asymptomatic, mild, moderate, severe and critical respectively. n = 1 for malignant (male).
Extended Data Fig. 9 Single-cell BCR analysis between genders across COVID-19 severity. a, Summary plot for BCR IGHV gene usage split by gender (top: female; bottom: male). Each data point is presented as the mean gene usage proportion + /- standard error of the mean for samples within each sample group. Kruskal-Wallis test with Benjamini-Hochberg *P < 0.05 (adjusted p = 0.035) n = 12, 5, 8, 20, 6 and 6 biologically independent male patients/samples and n = 12, 4, 15, 10, 7 and 5 biologically independent female patients/samples for healthy, asymptomatic, mild, moderate, severe and critical respectively. b, (Top) Scatter plot of clonotype/cluster size by vertex size gini indices computed from contracted BCR networks (identical nodes are merged and counted). Each dot represents the gini indices of an individual coloured by severity status. Marginal histograms indicate the distribution of samples in a given severity status along the axes. (Bottom, left) Cluster/clonotype size (contracted network) gini indices separated by gender. (Bottom, right) Vertex size (contracted network) gini indices separated by gender. The boxes extend from the 25th to 75th percentiles. The whiskers go down to the smallest value and up to the largest. The line in the middle of the boxes is plotted at the median. Statistical tests were performed with non-parametric Mann-Whitney U test between the gender groups within each severity status and were considered statistically significant if Benjamini-Hochberg corrected P < 0.05 (denoted by *; n.s. denotes not significant). The Benjamini-Hochberg adjusted p-values are as follows: cluster/clonotype size (contracted network) gini indices comparisons: 3.42x10-1, 1.91x10-25, 7.71x10-272, 1.31x10-169, 3.84x10-113 and 2.98x10-1; vertex size (contracted network) gini indices comparisons: 1.12x10-57, 2.17x10-1, <1x10-308, 3.64x10-49, 2.69x10-53 and 2.37x10-04 - for healthy, asymptomatic, mild, moderate, severe and critical respectively. Colour of asterisks indicates which gender group displays a higher mean gini index (yellow: female; grey: male). n = 12, 5, 8, 20, 6 and 6 biologically independent male patients/samples and n = 12, 4, 15, 10, 7 and 5 biologically independent female patients/samples for healthy, asymptomatic, mild, moderate, severe and critical respectively. n = 1 for malignant (male).CC BY · PubMed Central

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

Research works printed but not linked: 1

  1. World Health Organization. Clinical management of COVID-19. https://www.who.int/publications/i/item/clinical-management-of-covid-19 (2021). 35917394

Cited by

See all 805

Metadata

Authors
Cambridge Institute of Therapeutic Immunology and Infectious Disease-National Institute of Health Research (CITIID-NIHR) COVID-19 BioResource Collaboration, Stephenson, E., Reynolds, G., Botting, R. A., Calero-Nieto, F. J., Morgan, M. D., Tuong, Z. K., Bach, K., Sungnak, W., Worlock, K. B., Yoshida, M., Kumasaka, N., Kania, K., Engelbert, J., Olabi, B., Spegarova, J. S., Wilson, N. K., Mende, N., Jardine, L., Gardner, L. C. S., Goh, I., Horsfall, D., McGrath, J., Webb, S., Mather, M. W., Lindeboom, R. G. H., Dann, E., Huang, N., Polanski, K., Prigmore, E., Gothe, F., Scott, J., Payne, R. P., Baker, K. F., Hanrath, A. T., Schim van der Loeff, I. C. D., Barr, A. S., Sanchez-Gonzalez, A., Bergamaschi, L., Mescia, F., Barnes, J. L., Kilich, E., de Wilton, A., Saigal, A., Saleh, A., Janes, S. M., Smith, C. M., Gopee, N., Wilson, C., Coupland, P., Coxhead, J. M., Kiselev, V. Y., van Dongen, S., Bacardit, J., King, H. W., Rostron, A. J., Simpson, A. J., Hambleton, S., Laurenti, E., Lyons, P. A., Meyer, K. B., Nikolić, M. Z., Duncan, C. J. A., Smith, K. G. C., Teichmann, S. A., Clatworthy, M. R., Marioni, J. C., Göttgens, B., & Haniffa, M.
Affiliations
  • Newcastle University (GB)
  • European Bioinformatics Institute (GB)
  • University of Cambridge (GB)
  • Wellcome Sanger Institute (GB)
  • University College London (GB)
  • Wellcome/MRC Cambridge Stem Cell Institute (GB)
  • Ludwig-Maximilians-Universität München (DE)
  • Newcastle upon Tyne Hospitals NHS Foundation Trust (GB)
  • University College London Hospitals NHS Foundation Trust (GB)
  • Royal Free London NHS Foundation Trust (GB)
  • Queen Mary University of London (GB)
  • Sunderland Royal Hospital (GB)
Corresponding author
  • Teichmann, Sarah A.
  • Clatworthy, Menna R.
  • Marioni, John C.
  • Göttgens, Berthold
  • Haniffa, Muzlifah
Venue
Nature Medicine
Published
2021-04-20
Volume
27
Issue
5
Pages
904-916
ISSN
  • 1078-8956 (ISSN-L)
  • 1546-170X
OpenAlex
W3153732845
Cited by
742
References
81
Citations by year
2026: 84 · 2025: 115 · 2024: 148 · 2023: 153 · 2022: 166 · 2021: 74
FWCI
43.43
Type
journal article
Language
en
Open access
Licence
Locations
  • Nature Medicine (publishedVersion, cc-by)
  • medRxiv (acceptedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • Aberdeen University Research Archive (Aberdeen University) (submittedVersion)
  • UCL Discovery (University College London) (submittedVersion)
  • Institutional Research Information System (Università degli Studi di Brescia) (submittedVersion)
  • Institutional Research Information System (Università degli Studi di Brescia) (submittedVersion, other-oa)
  • Minerva Access (University of Melbourne) (submittedVersion, cc-by)
  • Queen Mary Research Online (Queen Mary University of London) (submittedVersion, cc-by)
  • Apollo (University of Cambridge) (submittedVersion, cc-by)
  • Repository for Publications and Research Data (ETH Zurich) (submittedVersion, cc-by)
  • Apollo (cc-by)
  • Apollo
  • ETH Zürich Research Collection (cc-by)
Funders
  • NHS Blood and Transplant
  • Wellcome Trust (204721/Z/16/Z, 206328/Z/17/Z, 207556/Z/17/Z, 211276/Z/18/Z, 214539/Z/18/Z, Z/17/Z, 107630/Z/15/Z, 212516/Z/18/Z, WT206194, 211153/Z/18/Z, WT107931/Z/15/Z, /Z/15/Z, Wellcome) (WT211276/Z/18/Z, WT107931/Z/15/Z, 213555/Z/18/Z, 207556_Z_17_Z, 211153/Z/18/Z, WT206194, 206328/Z/17/Z, 203151/Z/16/Z, 213555/Z/18/Z, WT211276/Z/18/Z, 203151)
  • European Molecular Biology Laboratory
  • UK Research and Innovation
  • Chan Zuckerberg Initiative (174169)
  • Action Medical Research (GN2779)
  • Cancer Research UK (A:29580, CRUK C9545/A29580, C9545, DRCPGM\100015, C9545/A29580, S_3517)
  • National Institute for Health and Care Research (CL-2016-01-001, NF-SI-0514-10109, CL‐2017‐01‐004, CL-2015-01-002, ACF-2018-01-004, ACF-2018-01-001, ACF-2020-01-007, CL-2013-01-003, RP-2017-08-ST2-002, NIHR201394)
  • Department of Health and Social Care
  • Academy of Medical Sciences (AMS-SGCL11-Duncan, SGL015\1021)
  • University College London
  • Newcastle University
  • Rosetrees Trust
  • Lister Institute of Preventive Medicine
  • Deutsche Forschungsgemeinschaft
  • Cambridge University Hospitals (BRC 2012-2017, 146281)
  • Jikei University School of Medicine
  • BMA Foundation for Medical Research
  • Versus Arthritis (21777)
  • UK Regenerative Medicine Platform
  • UK Coronavirus Immunology Consortium
  • NIHR BioResource
  • Medical Research Council (MR/S035842/1, MR/5005579/1, MR/P008801/1, MR/R015635/1, MR/K017047/1, MC_PC_17230, MR/S036113/1, MR/S005579/1, 203151/Z/16/Z)
  • Biotechnology and Biological Sciences Research Council (BB/V006738/1, BB/P002293/1)
  • NIHR Newcastle Biomedical Research Centre
  • NIHR Great Ormond Street Hospital Biomedical Research Centre
  • Health Data Research UK
Topics
  • Health Sciences › Medicine › Infectious Diseases › COVID-19 Clinical Research Studies
  • Health Sciences › Medicine › Infectious Diseases › SARS-CoV-2 and COVID-19 Research
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Single-cell and spatial transcriptomics
Keywords
COVID-19 pandemic, SARS-CoV-2, immune response, peripheral blood mononuclear cells, non-classical monocytes, complement system, platelet sequestration, alveolar macrophages, megakaryopoiesis, hematopoietic stem and progenitor cells, platelet activation, CD8+ T cells, T follicular helper cells, IgA2, plasmablasts, plasma cells, disease severity, single-cell multi-omics
MeSH terms
COVID-19, SARS-CoV-2, Cross-Sectional Studies, Humans, Monocytes, Receptors, Antigen, B-Cell, Receptors, Antigen, T-Cell, T-Lymphocytes, Proteome, Single-Cell Analysis, Transcriptome

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.