Comparing library preparation methods for SARS-CoV-2 multiplex amplicon sequencing on the Illumina MiSeq platform

Genomic surveillance has a key role in tracking the ongoing COVID-19 pandemic, but information on how different sequencing library preparation approaches affect the data produced are lacking. We compared three library preparation methods using both tagmentation (Nextera XT and Nextera Flex) and ligation-based (KAPA HyperPrep) approaches on both positive and negative samples to provide insights into any methodological differences between the methods, and validate their use in SARS-CoV-2 amplicon sequencing. We show that all three library preparation methods allow us to recover near-complete SARS-CoV-2 genomes with identical SNP calls. The Nextera Flex and KAPA library preparation methods gave better coverage than libraries prepared with Nextera XT, which required more reads to call the same number of genomic positions. The KAPA ligation-based approach shows the lowest levels of human contamination, but contaminating reads had no effect on the downstream analysis. We found some examples of library preparation-specific differences in minority variant calling. Overall our data shows that the choice of Illumina library preparation method has minimal effects on consensus base calling and downstream phylogenetic analysis, and suggests that all methods would be suitable for use if specific reagents are difficult to obtain.