Transcriptomics-based drug repositioning pipeline identifies 1 therapeutic candidates for COVID-19 2 3

Transcriptomics-based drug repositioning pipeline identifies 1 therapeutic candidates for COVID-19 2 3 Authors: Brian L. Le1,2+, Gaia Andreoletti1,2+, Tomiko Oskotsky1,2+, Albert Vallejo4 Gracia3, Romel Rosales4,5, Katharine Yu1,2,6, Idit Kosti1,2, Kristoffer E. Leon3, Daniel G. 5 Bunis1,2,6, Christine Li1,2,7, G. Renuka Kumar3, Kris M. White4,5, Adolfo García6 Sastre4,5,8,9, Melanie Ott3,10, Marina Sirota1,2* 7 8 Affiliations: 9 1 Department of Pediatrics, UCSF, SF, CA, USA 10 2 Bakar Computational Health Sciences Institute, UCSF, SF, CA, USA 11 3 Gladstone Institute of Virology, Gladstone Institutes, SF, CA, USA 12 4 Department of Microbiology, Icahn School of Medicine at Mount Sinai, New York, NY, 13 USA 14 5 Global Health and Emerging Pathogens Institute, Icahn School of Medicine at Mount 15 Sinai, New York, NY, USA 16 6 Biomedical Sciences Graduate Program, UCSF, SF, CA, USA 17 7 Shanghai American School, Shanghai, China 18 8 Department of Medicine, Division of Infectious Diseases, Icahn School of Medicine at 19 Mount Sinai, New York, NY, USA 20 9 The Tisch Cancer Institute, Icahn School of Medicine at Mount Sinai, New York, NY, 21 USA 22 10 Department of Medicine, UCSF, SF, CA, USA 23 24 + These authors contributed equally to this manuscript. 25 26 *Correspondence should be addressed to: marina.sirota@ucsf.edu 27 28


176
In this study, we applied our drug repositioning pipeline to SARS-CoV-2 differential 177 gene expression signatures derived from publicly available RNA-seq data ( Figure 1). The 178 transcriptomic data were generated from distinct types of tissues, so rather than 179 aggregating them together, we predicted therapeutics for each signature and then  these signatures can be found in the supplement (Tables S1, S2, S3).  After analyzing the input SARS-CoV-2 signatures, we utilized our repositioning 221 pipeline to identify drugs with reversed profiles from CMap ( Figure 1). Significantly 222 reversed drug profiles were identified for each of the signatures using a permutation 223 approach: 30 hits from the ALV signature (Table S4), 15 hits from the EXP signature 224 (Table S5), and 86 hits from the BALF signature (Table S6) (Table S7). Twenty-five drug hits 232 reversed at least two signatures (p = 0.0334, random sampling), and four drug hits 233 reversed three signatures (p = 0.0599, random sampling) ( Table 1, Figure 3A). 234 We further characterized the common hits by examining their interactions with 235 proteins in humans. We used known drug targets from DrugBank 32 and predicted 236 additional targets using the similarity ensemble approach (SEA) 33 . We visualized the 237 known interactions from DrugBank in a network. Figure 3B shows the connectivity across 238 compounds highlighting both single drug genes (such as SIGMAR1 for haloperidol) and 239 genes shared across drugs, such as ADRA2A and DRD1 for haloperidol and co-   can rapidly identify readily available potential therapeutics in antiviral contexts.

361
There are several limitations of our approach that should be recognized. In

368
In our work, we applied the KS-based similarity metric on the CMap database.

402
Study design 403 We have previously developed and used a transcriptomics based bioinformatics   (Table S1).

424
We additionally applied a fold-change cutoff (|log2 FC| > 2), resulting in 125 genes used 425 as the EXP signature (Table S2).  (Table S3).      We generated lists of statistically significant differentially expressed genes from the analysis of 775 three published studies of SARS-CoV-2 and COVID-19. The drug repositioning computational 776 pipeline compares the ranked differential expression of the COVID-19 disease signature with that 777 of drug profiles from CMap. A reversal score based on the Kolmogorov-Smirnov statistic is 778 generated for each disease-drug pair. If a drug profile significantly (FDR < 0.05) reverses the 779 disease signature, then the drug could be therapeutic for the disease. Across all datasets, a total 780 of 102 drugs have been identified as potentially therapeutic for COVID-19. Twenty-five drugs were 781 identified in analyses of at least two of the three datasets. We further conducted pathways 782 analyses and targeted analyses on the results, focusing on the 25 shared hits. Finally, we 783 validated sixteen of our top predicted hits in live SARS-CoV-2 antiviral assays. 784    The lack of a dose response in cell viability probably reflects cytostatic and not cytotoxic effects. 820 Data are mean ± s.d.; n = 3 biologically independent samples for cell viability data. 821