Respiratory viral co-infections among SARS-CoV-2 cases confirmed by virome capture sequencing

Abstract

Accumulating evidence supports the high prevalence of co-infections among Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) patients, and their potential to worsen the clinical outcome of COVID-19. However, there are few data on Southern Hemisphere populations, and most studies to date have investigated a narrow spectrum of viruses using targeted qRT-PCR. Here we assessed respiratory viral co-infections among SARS-CoV-2 patients in Australia, through respiratory virome characterization. Nasopharyngeal swabs of 92 SARS-CoV-2-positive cases were sequenced using pan-viral hybrid-capture and the Twist Respiratory Virus Panel. In total, 8% of cases were co-infected, with rhinovirus (6%) or influenzavirus (2%). Twist capture also achieved near-complete sequencing (> 90% coverage, > tenfold depth) of the SARS-CoV-2 genome in 95% of specimens with Ct < 30. Our results highlight the importance of assessing all pathogens in symptomatic patients, and the dual-functionality of Twist hybrid-capture, for SARS-CoV-2 whole-genome sequencing without amplicon generation and the simultaneous identification of viral co-infections with ease.

Figure 1

Viruses detected in SARS-CoV-2 case specimens by two hybrid-capture sequencing approaches. (a) VirCapSeq (n = 92); and (b) Twist Respiratory Virus Panel (n = 83). Heatmap of viral reads in log scale and represented at the genus level. Sample IDs apply for both panels (a) and (b), indicating overlapping samples sequenced by both approaches. Horizontal line separates respiratory viruses (above) from non-respiratory viruses (below). Bar charts indicate number of viruses detected per specimen. nCoV_neg_1 & 2 are clinical control specimens from two individuals confirmed negative for SARS-CoV-2 by qRT-PCR.

Figure 2

Full genome coverage of co-infecting influenzavirus. Coverage plot of sequence reads generated by Twist capture sequencing of the SARS-CoV-2 case specimen nCoV_240, aligned to the influenzavirus A reference genome across eight different segments (S1-S8). Depth represented as X fold coverage. Single nucleotide polymorphisms (SNPs) detected at positions across the genome are indicated in red, greater intensity of red indicates higher % frequency.

Figure 3

Complete SARS-CoV-2 genome coverage by sequences generated using the Twist Respiratory Virus Panel. (a) The number of sequence reads generated by VirCapSeq (blue) and Twist (red) hybrid-capture sequencing, aligned to SARS-CoV-2 in case samples (n = 83) with varying viral load determined by the qRT-PCR cycle threshold (Ct) value. (b) The fraction of SARS-CoV-2 genome covered at > 10X depth. Violin plot (left) shows the distribution of genome coverage in samples sequenced by VirCapSeq (blue) and Twist (red) capture with horizontal line indicating the median fraction of genome covered. Vertical dotted line indicates the Ct 30 border (right). (c) Distribution of aligned sequence reads across the SARS-CoV-2 reference genome (MN908947.3) and depth of coverage at each position in the genome (50-bp windows) normalized to the average coverage across the whole genome for a given sample.

Figure 4

Confirmation of a 328 nt ORF8 deletion in the SARS-CoV-2. Genome browser view of Twist enriched Illumina (upper) and amplicon-based WGS ONT (lower) sequencing reads aligned across the SARS-CoV-2 genome of nCoV_200 case specimen, zoomed in at the site of 328 nt ORF8 deletion. ONT sequence alignment shows loss of coverage in the region targeted by the A7 amplicon primers, due to the deletion of a primer-binding site within ORF8. In contrast, Twist sequence reads align continuously across this region (Genes orf3a to orf7), providing greater breadth of coverage of the reference genome (MN908947.3).