key: cord-0775881-88l4f63n authors: Lopez-Martinez, C.; Martin-Vicente, P.; Gomez de Ona, J.; Lopez-Alonso, I.; Gil-Pena, H.; Cuesta-Llavona, E.; Fernandez-Rodriguez, M.; Crespo, I.; Salgado del Riego, E.; Rodriguez-Garcia, R.; Parra, D.; Fernandez, J.; Rodriguez-Carrio, J.; Davalos, A.; Chapado, L. A.; Coto, E.; Albaiceta, G. M.; Amado-Rodriguez, L. title: Transcriptomic clustering of critically ill COVID-19 patients date: 2022-03-02 journal: nan DOI: 10.1101/2022.03.01.22271576 sha: a9b5299e2048f606e7944c8b4cbd6b11d22f1bb8 doc_id: 775881 cord_uid: 88l4f63n Infections caused by SARS-CoV-2 may cause a severe disease, termed COVID-19, with significant mortality. Host responses to this infection, mainly in terms of systemic inflammation, have emerged as key pathogenetic mechanisms, and their modulation is the only therapeutic strategy that has shown a mortality benefit. Herein, we used peripheral blood transcriptomes of critically-ill COVID-19 patients obtained at admission in an Intensive Care Unit (ICU), to identify two transcriptomic clusters characterized by expression of either interferon-related or immune checkpoint genes, respectively. These profiles have different ICU outcome, in spite of no major clinical differences at ICU admission. A transcriptomic signature was used to identify these clusters in an external validation cohort, yielding similar results. These findings reveal different underlying pathogenetic mechanisms and illustrate the potential of transcriptomics to identify patient endotypes in severe COVID-19, aimed to ultimately personalize their therapies. Instituto de Investigación Sanitaria del Principado de Asturias. Oviedo, Spain. 2 Centro de Investigación Biomédica en Red (CIBER)-Enfermedades respiratorias. Madrid, Spain. 3 Instituto Universitario de Oncología del Principado de Asturias. Oviedo, Spain. 4 Servicio de Genética Molecular. Hospital Universitario Central de Asturias. Oviedo, Spain. 5 Red de Investigación Renal (REDINREN), Madrid, Spain. 6 Departamento de Morfología y Biología Celular. Universidad de Oviedo. Oviedo, Spain. 7 Servicio de Pediatría. Hospital Universitario Central de Asturias. Oviedo, Spain. 8 Departamento de Biología Funcional. Universidad de Oviedo. Spain. 9 Unidad de Cuidados Intensivos Polivalente. Hospital Universitario Central de Asturias. Oviedo, Spain. 10 Unidad de Cuidados Intensivos Cardiológicos. Hospital Universitario Central de Asturias. Oviedo, Spain. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted March 2, 2022. ; https://doi.org/10.1101/2022.03.01.22271576 doi: medRxiv preprint Infections caused by SARS-CoV-2 have a wide range of severity, from asymptomatic to life-threatening cases. The most severe forms of Coronavirusinduced disease (termed COVID-19) 1 lead to respiratory failure fulfilling the acute respiratory distress syndrome (ARDS) criteria 2 . These critically ill patients often require mechanical ventilation and supportive therapy in an intensive care unit (ICU) and show mortality rates that range from 12 to 91% depending on patient and hospital factors 3 . Local and systemic inflammation are key pathogenetic mechanisms in severe COVID-19 4 . Viral infection triggers a host response that involves not only antiviral mechanisms, such as release of interferons, but may also activate a systemic, non-specific inflammatory response that has been related to multiple organ failure and death 5 . The only treatments that have shown a survival benefit in critically-ill COVID-19 patients aim to modulate this inflammatory response 6 . However, it has been suggested that these treatments do not benefit patients with less severe forms of the disease or with only a mild activation of inflammation 7, 8 . There is increasing evidence that ARDS patients show different clinical features or systemic responses to severe diseases (phenotypes and endotypes respectively) 9 . Although the underlying causes responsible for this heterogeneity are not fully understood, clinical data showing different outcomes in response to a given treatment suggest that pathogenetic mechanisms may be different 10 . Therefore, identification of patient pheno/endotypes may be relevant not only for risk stratification, but also to design specific, personalized therapies in the ICU. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted March 2, 2022 Transcriptomic profiling after sequencing of whole blood RNA may be useful to identify groups of critically-ill patients with different underlying pathogenetic mechanisms [13] [14] [15] . In addition, preliminary results suggest that micro-RNA (miRNA) expression could also play a role in this setting 16 . We hypothesized that clustering of COVID-19 patients using transcriptomics at ICU admission could help to identify subgroups with different pathogenesis. To test this hypothesis, we prospectively sequenced peripheral blood RNA and serum miRNA at ICU admission in a cohort of COVID-19 patients, applied an unbiased clustering algorithm and compared gene expression clinical data and outcomes in the identified subgroups. Finally, we validated our findings in an external cohort. This prospective observational study was reviewed and approved by the regional ethics committee (Comité de Ética de la Investigación Clínica del Principado de Asturias, ref 2020.188). Informed consent was obtained from each patient's next of kin. Fifty-six consecutive patients admitted to one of the participant ICUs at Hospital Universitario Central de Asturias (Oviedo, Spain) from April to December 2020 were included in the study. Inclusion criteria were . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) After inclusion, two samples of peripheral blood were drawn in the first 72 hours after ICU admission. One sample was collected in Tempus Blood RNA tubes (Thermo Fisher) to facilitate cell lysis, precipitate RNA and prevent its degradation. The other sample was immediately centrifuged to obtain serum and mixed with TRI reagent for serum RNA precipitation. These tubes were stored at -80°C until processing. Whole blood RNA was extracted by isopropanol precipitation and sequenced in an Ion S5 GeneStudio sequencer using AmpliSeq Transcriptome Human Gene Expression kits that amplify all the canonical human transcripts. Details on RNA extraction and sequencing have been provided elsewhere 8 . FASTQ files containing RNA sequences were pseudoaligned using a reference transcriptome (http://refgenomes.databio.org) and salmon software 17 to obtain transcript counts. Total serum RNA was extracted using miRNEasy kit (Qiagen), following manufacturer's instructions, and miRNAs isolated and sequenced at BGI Genomics (Wuhan, China). miRNA readouts were mapped using bowtie2 18 , with an index built using the hg38 human reference genome. Quantification of sequenced miRNAs was performed using miRDeep2 19 with reference human . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) Clustering of RNA samples was performed following a previously described protocol 20 . Briefly, log 2 -transformed gene expression data (expressed as transcripts per million reads) were filtered to keep the 5% of features with the largest variance. Clusters were built based on Euclidean distances following the Ward clustering algorithm. Cluster p-values, indicating how strong the cluster is supported by the data, were calculated by multiscale bootstrap resampling using the pvclust package 21 for R. Gene raw counts obtained after pseudoalignment were compared between clusters using DESeq2 22 . Log 2 fold change for each gene between variants and the corresponding adjusted p-value (corrected using a false discovery rate of 0.05) were calculated. Genes with an absolute log 2 fold change above 2 and an adjusted p-value lower than 0.01 were used for Gene Set Enrichment Analysis (GSEA) using the clusterProfiler R package 23 . A correlation analysis was performed in genes annotated to a Gene Ontology category involved in interferon pathway. Correlation coefficients between each gene pair were transformed to z-scores and the p-values for each comparison calculated using the DGCA package for R 24 . Genes with opposite correlations in each cluster were selected and the networks defined by their significant correlations traced. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted March 2, 2022. ; https://doi.org/10.1101/2022.03.01.22271576 doi: medRxiv preprint Differentially expressed genes between clusters were also matched with the c-miRNAs expressed for each group using the MicroRNA Target Filter tool from Ingenuity Pathway Analysis (Qiagen Digital Insights), to identify predicted interactions. Intersected mRNA and miRNA datasets were filtered to explicitly pair opposed and reciprocal expression changes. Only experimentally observed predictions were considered. Key mRNA-miRNA relationships identified were overlayed onto the networks of interest to explore the predicted functionality in our datasets. Pathways related to humoral, and T and B cellular immune responses were selected as relevant. miRNAs with <3 targeted mRNAs were filtered out from the network. Demographics and comorbidities were collected at ICU admission (day 1). Data on gas exchange, respiratory support, hemodynamics, received treatments and results from routine laboratory analyses were prospectively collected at days 1 and 7 after ICU admission. Patients were followed up to ICU discharge. During this period, duration of ventilatory support and vital status were collected for outcome analysis. Proportions of transcriptionally active circulating cells in each sample were estimated using Immunostates 25 , a previously published deconvolution algorithm. From the original reference matrix, cell populations not commonly identified in peripheral blood (Mast cells and macrophages) were removed. Using this modified reference matrix containing expression of 318 genes for 16 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted March 2, 2022. ; https://doi.org/10.1101/2022.03.01.22271576 doi: medRxiv preprint different blood cell types, the percentage of each one of these types was estimated from the bulk RNAseq. To validate our results in an external cohort, we used a publicly available dataset of 50 transcriptomes from critically-ill COVID-19 patients 26 were compared between clusters. Data are expressed as median and interquartile range. Missing data were not imputed. Differences between clusters were assessed using two-tailed Wilcoxon or chi-square tests (for quantitative and qualitative data respectively). For survival analysis, patients were followed up to ICU discharge, with ICU discharge alive and spontaneously breathing being the main outcome . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted March 2, 2022. ; https://doi.org/10.1101/2022.03.01.22271576 doi: medRxiv preprint 1 0 measurement. Differences in this outcome between clusters were assessed using a competing risk model as previously described 8 , and hazard ratio for the main outcome, with the corresponding 95% confidence interval, was calculated. All the analyses were performed using R v4.1.1 27 and packages ggplot2 28 , pROC 29 and survival 30 , in addition to those previously cited. All the code and raw data can be found at https://github.com/Crit-Lab/COVID_clustering. Peripheral Therefore, the sample was divided in two mutually exclusive groups, termed COVID-19 transcriptomic profiles (CTP) 1 and 2. Bidimensional representation of the study population using a UMAP algorithm confirmed the separation of the two clusters ( Figure 1C ). Supplementary figure 2 shows a heatmap with the expression of the genes used for clustering. Then we assessed the overall differences in gene expression. Using an adjusted p-value cut-off point of 0.01, there were 9700 differentially expressed genes (Supplementary file 1), with 3640 having an absolute log 2 fold change above 2 ( Figure 2A ). Interestingly, most of these genes were downregulated in . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) 1 1 CTP2. Then, GSEA was used to identify the molecular pathways involving these differentially expressed genes. One hundred and ten biological processes with significant differences between clusters were identified (Supplementary Figure 2D ) and regulatory T-cell differentiation ( Figure 2E ). In addition to these quantitative changes in expression of interferon-related genes, we explored the existence of qualitative differences between clusters. We calculated the linear correlation coefficients among the 145 genes included in the Gene ontology categories involving interferon signaling in each cluster. There was a significant difference between the two correlation matrices ( Figure 3 , p<0.001 calculated using a Chi-square test), thus demonstrating differences in the orchestration/structure of IFN responses between groups. In addition, pairwise differences in correlation coefficients for each gene pair were assessed. Gene pairs with correlation coefficients with an adjusted p-value for their difference below 0.05 and opposite signs in each cluster were selected, and networks including these genes traced ( Figure 3 and Supplementary Figure 4 ). These results suggest that both clusters have a qualitatively different activation of the interferon pathway, with some genes such as HSP90AB1 and JAK1 acting as hubs with opposite correlations. Of note, CTP1 was hallmarked . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted March 2, 2022. ; https://doi.org/10.1101/2022.03.01.22271576 doi: medRxiv preprint by strong, positive correlations among effector IFN proteins, whereas this was not the case for CTP2. The previous results suggest that the identified clusters may have a different circulating lymphocyte profile. To further explore this finding, cell populations were estimated by deconvolution of RNAseq data. This analysis revealed a higher granulocyte proportion in patients assigned to CTP1, a lower proportion of lymphocytes and no differences in monocytes or NK cells ( Figure To identify miRNAs potentially related to the observed changes in RNA expression, we analyzed miRNA content using the Mirna Target Filter included in Ingenuity Pathway Analysis. After filtering by experimentally confirmed miRNA-gene relationships, and only opposed changes in miRNA/gene expression levels, 83 miRNAs targeting 608 genes were identified in our dataset. Given the observed differences in lymphocyte populations, we focused on miRNAs involved in humoral and cellular immune regulation (29 miRNAs . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. miRNAs predicted to regulate expression of these genes were identified and compared ( Figure 5B -H). Among these, only counts of miR-145a-5p and miR-181-5p were significatively lower in CTP2 ( Figure 5C and 5D respectively). Clinical differences between clusters at ICU admission were studied (Table 1 ). There were no significant differences in demographic and clinical variables other than a higher leukocyte count in cluster CTP1, with no differences in lymphocyte counts. Patients assigned to CTP2 cluster showed more ventilatorfree days during the first 28 days in ICU (Table 1 ). In the survival analysis, after adjusting for age, sex, and need for intubation during the ICU stay, assignation to CTP2 increased the probability of ICU discharge alive and spontaneously breathing (HR 2.00 [1.08 -3.70], p=0.028, Figure 6 ). To apply our findings to an external cohort, we first developed a characteristic gene signature that allows assignation to one cluster using gene expression data. We focused on genes upregulated in CTP2, as they constitute a relatively small group, given the massive gene downregulation in this group. Among these 117 upregulated genes, 15 (BCL2, CARD11, CD247. CD7, CD81, CLSTN1, E2F6, MCM5, PARP1, PNPO, RASGRP1, RCC2, RPTOR, RUNX3 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted March 2, 2022. ; https://doi.org/10.1101/2022.03.01.22271576 doi: medRxiv preprint and ZAP70) had an AUROC to identify CTP2 higher than 0.95. Expression of these genes was synthesized into a transcriptomic score. As expected, the score was higher in CTP2 (Supplementary figure 7A) , with an AUROC of 0.99 (95% CI 0.97 -1) (Supplementary figure 7B) . In a Cox-regression analysis including this transcriptomic score, age, sex and need for mechanical ventilation, the score was correlated to ICU discharge (HR 1.002 [1.000 -1.003], p=0.012). Based on these results, a cut-off point of 250 in this score, aimed to include all CTP2 cases, was chosen. Then, this transcriptomic score was calculated in an external cohort of 50 severe COVID-19 patients with publicly available blood gene expression in samples obtained at enrolment. After computing transcriptomic scores, 13 patients were classified as CTP1 and 37 as CTP2. Comparisons between these clusters are shown in Table 2 . In spite of no significant differences in age, sex, APACHE-II or SOFA scores, patients assigned to CTP2 showed more ventilator-free days at day 28 of ICU stay, and the percentage of patients with zero ventilator-free days at day 28 was lower in CTP2. Deconvolution of peripheral blood transcriptomes in this validation cohort recapitulated some of the differences observed in the discovery cohort, including higher neutrophil counts and lower proportions of CD8+ T-cells in CTP1 (Supplementary Figure 8 ). Our results show that unsupervised clustering of critically ill COVID-19 patients, using transcriptomic profiles from peripheral blood obtained at ICU admission, results in two groups with a differential immune response and outcome. In spite . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted March 2, 2022. ; https://doi.org/10.1101/2022.03.01.22271576 doi: medRxiv preprint of no clinical differences at admission other than the absolute leukocyte count, the cluster of patients with an enriched interferon response shows lower ICU survival rates. Application of a cluster-specific score to an independent cohort confirmed this result. These findings suggest that there are specific COVID endotypes with different underlying immunopathogenesis and outcomes. Clustering strategies have been proposed to identify different subgroups of critically ill patients with respiratory failure, that may help to personalize treatments. Two different phenotypes have been identified in several cohorts using clinical and laboratory data 9, 31 . A hyperinflammatory/reactive phenotype, characterized by higher concentrations of markers of acute inflammation, such as IL-6, IL-8, C-reactive protein, and tissue hypoxia has been linked to higher mortality rates and could specifically benefit from fluid restriction, higher PEEP levels or protective ventilation 32 , in contraposition to the uninflamed phenotype. Of note, causes of ARDS were different between the two clusters, with a higher incidence of sepsis in the inflammatory/reactive group. Opposed to a syndromic approach, clustering within a specific disease such as COVID-19 using routine clinical data has yielded conflicting results. Whereas direct translation of the inflammatory/reactive framework to a single-center cohort has identified equivalent groups 11 , other multicenter studies failed to identify COVID-19 subgroups at ICU admission 12 . In this setting, transcriptomic clustering may offer several advantages by including a large number of features for classification, reduced intervention times and absence of imputed or not available data. Bulk peripheral blood RNAseq has been used to identify relevant pathogenetic mechanisms in COVID-19, by comparing cases with different severity or against healthy . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. 38, 39 . Although an enhanced interferon-mediated response may be detrimental, it must be noted that loss-of-function variants of genes from the interferon pathway (such as IRF7 or IFNAR1) or autoantibodies . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. In summary, our results show that transcriptomic clustering using peripheral blood RNA at ICU admission allows the identification of two groups of criticallyill COVID-19 with different immune profile and outcome. These findings could . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted March 2, 2022. ; https://doi.org/10.1101/2022.03.01.22271576 doi: medRxiv preprint be useful for risk stratification of these patients and help to identify specific profiles that could benefit from personalized treatments aimed to modulate the inflammatory response or its consequences. The authors want to thank all the personnel at the participating ICUs and laboratories for their support during the development of the study. This work is . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted March 2, 2022. ; https://doi.org/10.1101/2022.03.01.22271576 doi: medRxiv preprint The authors declare no competing interests. All data and code used in the manuscript have been deposited in Git-Hub Correspondence and requests to Laura Amado-Rodríguez (lar@crit-lab.org) or Guillermo M Albaiceta (gma@crit-lab.org). . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted March 2, 2022. ; https://doi.org/10.1101/2022.03.01.22271576 doi: medRxiv preprint . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted March 2, 2022. ; https://doi.org/10.1101/2022.03.01.22271576 doi: medRxiv preprint . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted March 2, 2022. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted March 2, 2022. ; https://doi.org/10.1101/2022.03.01.22271576 doi: medRxiv preprint . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted March 2, 2022. LAG3 LAG3 LAG3 LAG3 LAG3 LAG3 LAG3 LAG3 LAG3 LAG3 LAG3 LAG3 LAG3 LAG3 LAG3 LAG3 TNFRSF18 TNFRSF18 TNFRSF18 TNFRSF18 TNFRSF18 TNFRSF18 TNFRSF18 TNFRSF18 TNFRSF18 TNFRSF18 TNFRSF18 TNFRSF18 TNFRSF18 TNFRSF18 TNFRSF18 TNFRSF18 TNFRSF18 LILRB4 LILRB4 LILRB4 LILRB4 LILRB4 LILRB4 LILRB4 LILRB4 LILRB4 LILRB4 LILRB4 LILRB4 LILRB4 LILRB4 LILRB4 IFNA6 IFNA6 IFNA6 IFNA6 IFNA6 IFNA6 IFNA6 IFNA6 IFNA6 IFNA6 IFNA6 IFNA6 IFNA6 IFNA6 IFNA6 IFNA6 IFNA6 IFNA2 IFNA2 IFNA2 IFNA2 IFNA2 IFNA2 IFNA2 IFNA2 IFNA2 IFNA2 IFNA2 IFNA2 IFNA2 IFNA2 IFNA2 IFNA2 IFNA2 IFNA17 IFNA17 IFNA17 IFNA17 IFNA17 IFNA17 IFNA17 IFNA17 IFNA17 IFNA17 IFNA17 IFNA17 IFNA17 IFNA17 IFNA17 IFNA17 IFNA17 IFNA10 IFNA10 IFNA10 IFNA10 IFNA10 IFNA10 IFNA10 IFNA10 IFNA10 IFNA10 IFNA10 IFNA10 IFNA10 IFNA10 IFNA10 IFNA10 IFNA10 IFNA14 IFNA14 IFNA14 IFNA14 IFNA14 IFNA14 IFNA14 IFNA14 IFNA14 IFNA14 IFNA14 IFNA14 IFNA14 IFNA14 IFNA14 IFNA14 IFNA14 IFNA13 IFNA13 IFNA13 IFNA13 IFNA13 IFNA13 IFNA13 IFNA13 IFNA13 IFNA13 IFNA13 IFNA13 IFNA13 IFNA13 IFNA13 IFNA13 IFNA13 IFNA21 IFNA21 IFNA21 IFNA21 IFNA21 IFNA21 IFNA21 IFNA21 IFNA21 IFNA21 IFNA21 IFNA21 IFNA21 IFNA21 IFNA21 IFNA21 IFNA21 IFNA4 IFNA4 IFNA4 IFNA4 IFNA4 IFNA4 IFNA4 IFNA4 IFNA4 IFNA4 IFNA4 IFNA4 IFNA4 IFNA4 IFNA4 IFNA4 IFNA4 IFNB1 IFNB1 IFNB1 IFNB1 IFNB1 IFNB1 IFNB1 IFNB1 IFNB1 IFNB1 IFNB1 IFNB1 IFNB1 IFNB1 IFNB1 IFNB1 IFNB1 IFNA8 IFNA8 IFNA8 IFNA8 IFNA8 IFNA8 IFNA8 IFNA8 IFNA8 IFNA8 IFNA8 IFNA8 IFNA8 IFNA8 IFNA8 IFNA8 IFNA8 IFNA7 IFNA7 IFNA7 IFNA7 IFNA7 IFNA7 IFNA7 IFNA7 IFNA7 IFNA7 IFNA7 IFNA7 IFNA7 IFNA7 IFNA7 IFNA7 IFNA7 IFNA16 IFNA16 IFNA16 IFNA16 IFNA16 IFNA16 IFNA16 IFNA16 IFNA16 IFNA16 IFNA16 IFNA16 IFNA16 IFNA16 IFNA16 IFNA16 IFNA16 IFNA1 IFNA1 IFNA1 IFNA1 IFNA1 IFNA1 IFNA1 IFNA1 IFNA1 IFNA1 IFNA1 IFNA1 IFNA1 IFNA1 IFNA1 IFNA1 IFNE IFNE IFNE IFNE IFNE IFNE IFNE IFNE IFNE IFNE IFNE IFNE IFNE IFNE IFNE IFNE IFNE CARD9 CARD9 CARD9 CARD9 CARD9 CARD9 CARD9 CARD9 CARD9 CARD9 CARD9 CARD9 CARD9 CARD9 CARD9 CARD9 CARD9 IFNL1 IFNL1 IFNL1 IFNL1 IFNL1 IFNL1 IFNL1 IFNL1 IFNL1 IFNL1 IFNL1 IFNL1 IFNL1 IFNL1 IFNL1 IFNL1 IFNL1 CCL19 CCL19 CCL19 CCL19 CCL19 CCL19 CCL19 CCL19 CCL19 CCL19 CCL19 CCL19 CCL19 CCL19 CCL19 CCL19 CCL19 IL5 IL5 IL5 IL5 IL5 IL5 IL5 IL5 IL5 IL5 IL5 IL5 IL5 IL5 IL5 IL5 BCL2 BCL2 BCL2 BCL2 BCL2 BCL2 BCL2 BCL2 BCL2 BCL2 BCL2 BCL2 BCL2 BCL2 BCL2 BCL2 BCL2 IGHA1 IGHA1 IGHA1 IGHA1 IGHA1 IGHA1 IGHA1 IGHA1 IGHA1 IGHA1 IGHA1 IGHA1 IGHA1 IGHA1 IGHA1 IGHA1 IGHA1 IGHG1 IGHG1 IGHG1 IGHG1 IGHG1 IGHG1 IGHG1 IGHG1 IGHG1 IGHG1 IGHG1 IGHG1 IGHG1 IGHG1 IGHG1 IGHG1 IGHG1 GCSAML GCSAML GCSAML GCSAML GCSAML GCSAML GCSAML GCSAML GCSAML GCSAML GCSAML GCSAML GCSAML GCSAML GCSAML IRF3 SHMT2 IFNA21 IFNW1 CH25H RNF26 IFNA10 IFNA6 IFNA2 IFNA17 IFNE IFNA14 IFNA1 IFNA13 SMPD1 IFNA4 UBE2K DHX33 TTLL12 UBE2G2 HSP90AB1 YTHDF2 FADD TPR JAK1 MUL1 PTPN1 CYLD ADAR NFKB1 MYD88 POLA1 PTPN11 IRF3 SHMT2 IFNA21 IFNW1 CH25H RNF26 IFNA10 IFNA6 IFNA2 IFNA17 IFNE IFNA14 IFNA1 IFNA13 SMPD1 IFNA4 UBE2K DHX33 TTLL12 UBE2G2 HSP90AB1 YTHDF2 FADD TPR JAK1 MUL1 PTPN1 CYLD ADAR NFKB1 MYD88 POLA1 PTPN11 IRF3 SHMT2 IFNA21 IFNW1 CH25H RNF26 IFNA10 IFNA6 IFNA2 IFNA17 IFNE IFNA14 IFNA1 IFNA13 SMPD1 IFNA4 UBE2K DHX33 TTLL12 UBE2G2 HSP90AB1 YTHDF2 FADD TPR JAK1 MUL1 PTPN1 CYLD ADAR NFKB1 MYD88 POLA1 PTPN11 IRF3 SHMT2 IFNA21 IFNW1 CH25H RNF26 IFNA10 IFNA6 IFNA2 IFNA17 IFNE IFNA14 IFNA1 IFNA13 SMPD1 IFNA4 UBE2K DHX33 TTLL12 UBE2G2 HSP90AB1 YTHDF2 FADD TPR JAK1 MUL1 PTPN1 CYLD ADAR NFKB1 MYD88 HSP90AB1 CYLD IFNA21 IFNA10 YTHDF2 IFNA1 IFNE IFNA13 IFNA14 IFNA4 IFNA2 CH25H IFNW1 IFNA6 IFNA17 HSP90AB1 CYLD IFNA21 IFNA10 YTHDF2 IFNA1 IFNE IFNA13 IFNA14 IFNA4 IFNA2 CH25H IFNW1 IFNA6 IFNA17 MUL1 DHX33 JAK1 IRF3 TTLL12 Risk Factors Associated With Mortality Among Patients With COVID-19 in Intensive Care Units in Lombardy, Italy Acute respiratory distress syndrome: the Berlin Definition Hospital-Level Variation in Death for Critically Ill Patients with COVID-19 The trinity of COVID-19: immunity, inflammation and intervention Immune dysregulation and immunopathology induced by SARS-CoV-2 and related coronaviruses -are we our own worst enemy? Clinical management of COVID-19 patients: living guideline Dexamethasone in Hospitalized Patients with Covid-19 Effects of IFIH1 rs1990760 variants on systemic inflammation and outcome in critically ill COVID-19 patients in an observational translational study Subphenotypes in acute respiratory distress syndrome: latent class analysis of data from two randomised controlled trials Acute respiratory distress syndrome subphenotypes and differential response to simvastatin: secondary analysis of a randomised controlled trial Latent Class Analysis Reveals COVID-19-related Acute Respiratory Distress Syndrome Subgroups with Differential Responses to Corticosteroids Longitudinal respiratory subphenotypes in patients with COVID-19-related acute respiratory distress syndrome: results from three observational cohorts Classification of patients with sepsis according to blood genomic endotype: a prospective cohort study Predicting sepsis severity at first clinical presentation: The role of endotypes and mechanistic signatures Genomic landscape of the individual host response and outcomes in sepsis: a prospective cohort study Circulating microRNA profiles predict the severity of COVID-19 in hospitalized patients Salmon provides fast and bias-aware quantification of transcript expression Fast gapped-read alignment with Bowtie N. miRDeep2 accurately identifies known and hundreds of novel microRNA genes in seven animal clades Clustering of RNA-Seq samples: Comparison study on cancer data Hierarchical Clustering with P-Values via Multiscale Bootstrap Resampling Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2 clusterProfiler 4.0: A universal enrichment tool for interpreting omics data DGCA: A comprehensive R package for Differential Gene Correlation Analysis Leveraging heterogeneity across multiple datasets increases cell-mixture deconvolution accuracy and reduces biological and technical biases Large-Scale Multi-omic Analysis of COVID-19 R: A language and environment for statistical computing Elegant Graphics for Data Analysis pROC: an open-source package for R and S+ to analyze and compare ROC curves A Package for Survival Analysis in R Understanding Heterogeneity in Biologic Phenotypes of Acute Respiratory Distress Syndrome by Leukocyte Expression Profiles Biotrauma during ultra-low tidal volume ventilation and venoarterial extracorporeal membrane oxygenation in cardiogenic shock: a randomized crossover clinical trial Monocyte-driven atypical cytokine storm and aberrant neutrophil activation as key mediators of COVID-19 disease severity Deciphering the state of immune silence in fatal COVID-19 patients Asymptomatic COVID-19: disease tolerance with efficient anti-viral immunity against SARS-CoV-2 Genetic mechanisms of critical illness in Covid-19 Severe Covid-19 GWAS Group et al. Genomewide Association Study of Severe Covid-19 with Respiratory Failure Increased CD4/CD8 ratio as a risk factor for critical illness in coronavirus disease 2019 (COVID-19): a retrospective multicentre study