Fig. 5

Quality inspection of the SARS-CoV-2 proteomics dataset. a Inspection revealed a separation of samples along the second principal component likely related to a plating effect. This was compensated for in subsequent statistical tests by including it as a covariate. Separation along the first axis is related to infection and time passed from infection. b The impact of performing differential expression analysis without and with inclusion of the putative plating number as a covariate. The inclusion of the covariate yielded 345 new statistical features while losing six as compared to not including the covariate. c Comparison of control samples 6 h and 2 h shows many features with a decrease in abundance, indicating that the mock treatment might influence the data. Comparison between infected samples at 6 h and 2 h show more limited differences, with seven detected viral proteins among those with increased abundance at 6 h indicated in red circles (log2 fold change > 0.3) out of which two passed the FDR threshold