SARS-CoV-2 exhibits intra-host genomic plasticity and low-frequency polymorphic quasispecies

Timokratis Karamitros; Gethsimani Papadopoulou; Maria Bousali; Anastasios Mexias; Sotiris Tsiodras; Andreas Mentis

In this study we analysed NGS data derived from clinical specimens (oral swabs) from three Chinese patients infected by SARS-CoV-2 (SRA projects PRJNA601736 and PRJNA603194). We aligned the raw read data on reference strain MN975262.1 using bowtie2 [18] , after quality check with FastQC (www.bioinformatics.bbsrc.ac.uk/projects/fastqc). The resulting alignments were visualized with the Integrated Genomics Viewer (IGV) [19] . After removing PCR duplicates, SNVs were called with a Bonferroni-corrected P-value threshold of 0.05 using samtools [20] and LoFreq [21] . Variants supported by absolute read concordance (>98%) were filtered-out from intra-host variant frequency calculations. We annotated the variations to the reference strain using snpEff [22] , SNVs effects were further filtered with snpSift [23] and we estimated the average mutation rate per gene across the viral genome using R scripts. We compared the localization of the intra-host SNVs with all available SNVs observed at population level up to February 18 th 2020 (retrieved from www.GISAID.org). We also compared all intra-host and population level SNPs with all primers and probes coordinates to investigate for potential interferences with currently available molecular diagnostic assays [24] (www.who.int/docs/default-source/coronaviruse/peiris-protocol-16-1-20.pdf).