A comparison of high-throughput SARS-CoV-2 sequencing methods from nasopharyngeal samples

Abstract

The COVID-19 pandemic caused by the new Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) continues to threaten public health and burden healthcare systems worldwide. Whole SARS-CoV-2 genome sequencing has become essential for epidemiological monitoring and identification of new variants, which could represent a risk of increased transmissibility, virulence, or resistance to vaccines or treatment. Different next-generation sequencing approaches are used in SARS-CoV-2 sequencing, although with different ability to provide whole genome coverage without gaps and to reliably detect new variants. In this study, we compared the performance of three target enrichment methods (two multiplex amplification methods and one hybridization capture) using nasopharyngeal swabs from infected individuals. We applied these target enrichment methods to the same set of nasopharyngeal samples (N = 93) in high-throughput mode. SARS-CoV-2 genome was obtained using short-read next-generation sequencing. We observed that each method has some advantages, such as high mapping rate (CleanPlex and COVIDSeq) or absence of systematic variant calling error (SureSelect) as well as their limitations such as suboptimal uniformity of coverage (CleanPlex), high cost (SureSelect) or supply shortages (COVIDSeq). Nevertheless, each of the three target enrichment kits tested in this study yielded acceptable results of whole SARS-CoV-2 genome sequencing and either of them can therefore be used in prospective programs of genomic surveillance of SARS-CoV-2. Genomic surveillance will be crucial to overcoming the ongoing pandemic of COVID-19, despite its successive waves and continually emerging variants.

Figure 1

Comparison of target enrichment kits performance. (a) SARS-CoV-2-specific mapping rate. (b) Breadth of coverage of the SARS-CoV-2 genome, where the depth of coverage is at least 10× . (c) Median depth of coverage.

Figure 2

Depth of coverage profile of a typical library constructed with different kits.

Figure 3

Departure from optimal depth of coverage. Relative depth was determined by comparing per-base depth to the mean depth of each library. Horizontal grey lines mark the optimal range between half the mean and twice the mean value (displayed on the right side of the y-axis).

Figure 4

Comparison of variant calling profile of the same individual (lineage B.1.160). Vertical colored lines represent variants called; the absence of lines indicates a match with the reference sequence (accession no. NC_045512.2). Black arrows at the top show variant calls that differ among the three methods (in this case nt positions C7390T and G19518T).

Figure 5

Venn diagram with the number of SNP sites called by each method. Among the 504 total SNPs sites observed in this study, 23 were not called consistently by all three methods.