Host, Viral, and Environmental Transcriptome Profiles of the Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2)

Daniel J Butler; Christopher Mozsary; Cem Meydan; David C Danko; Jonathan Foox; Joel Rosiene; Alon Shaiber; Ebrahim Afshinnekoo; Matthew MacKay; Fritz J Sedlazeck; Nikolay A Ivanov; Maria A Sierra; Diana Pohle; Michael Zeitz; Vijendra Ramlall; Undina Gisladottir; Craig D Westover; Krista Ryon; Benjamin Young; Chandrima Bhattacharya; Phyllis Ruggiero; Bradley W Langhorst; Nathan A Tanner; Justyn Gawrys; Dmitry Meleshko; Dong Xu; Jenny Xiang; Angelika Iftner; Daniela Bezdan; John Sipley; Lin Cong; Arryn Craney; Priya Velu; Ari Melnick; Iman A Hajirasouliha; Thomas Iftner; Mirella Salvatore; Massimo Loda; Lars F Westblade; Shawn Levy; Melissa Cushing; Nicholas P Tatonetti; Marcin Imielinski; Hanna Rennert; Christopher Mason

The reads that mapped unambiguously to the human reference genome via Kraken2 were used to detect the host transcriptional response to the virus. Reads matching Homo sapiens were trimmed with TrimGalore, aligned with STAR (v2.6.1d) to the human reference build GRCh38 and the GENCODE v33 transcriptome reference, gene expression was quantified using featureCounts, stringTie and salmon using the nf-core RNAseq pipeline (Pertea et al., 2015; Malinen et al., 2005; Johnson et al., 2007; Robinson et al., 2010; Naccache et al., 2014; Zamani et al., 2017; Ewels et al., 2019) . Sample QC was reported using fastqc, RSeQC, qualimap, dupradar, Preseq and MultiQC (Okonechnikov et al., 2016; Andrews, 2015; Ewesl et al., 2016; Sayols et al., 2016; Wang et al., 2012) . Samples that had more than 10 million human mapped reads were used for differential expression analysis. Reads, as reported by featureCounts, were normalized using variance-stabilizing transform (vst) in DESeq2 package in R (Love et al., 2014) for visualization purposes in log-scale. DESeq2 was used to call differential expression with either Positive cases vs Negative, or viral load (High/Medium/Low/None) as reported by either qRT-PCR cycle threshold (Ct) values or the combination viral load method as explained before. Genes with BH-adjusted p-value < 0.01 and absolute log2 fold-change greater than 0.58 (at least 50% change in either direction) were taken as significantly differentially regulated (Benjamini and Hochberg, 1995) . The complete gene list for all comparisons are given in Supp Table 3 . Resulting gene sets were ranked using log2 fold-change values within each comparison and put into GSEA to calculate gene set enrichment for molecular signatures database (MSigDB), MGI Mammalian Phenotypes database and ENCODE transcription factor binding sets (Liberzon et al., 2011; Subramanian et al., 2005; Sergushichev, 2016; Smith et al., 2018) . Any signature with adjusted p-value < 0.01 and absolute normalized enrichment score (NES) >= 1.5 were reported (Supp Table 3 ). author/funder. All rights reserved. No reuse allowed without permission.