SARS-CoV-2 Genome Sequencing Methods Differ in Their Abilities To Detect Variants from Low-Viral-Load Samples

Abstract

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) genomic surveillance has been vital in understanding the spread of coronavirus disease 2019 (COVID-19), the emergence of viral escape mutants, and variants of concern. However, low viral loads in clinical specimens affect variant calling for phylogenetic analyses and detection of low-frequency variants, important in uncovering infection transmission chains. ABSTRACT Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) genomic surveillance has been vital in understanding the spread of coronavirus disease 2019 (COVID-19), the emergence of viral escape mutants, and variants of concern. However, low viral loads in clinical specimens affect variant calling for phylogenetic analyses and detection of low-frequency variants, important in uncovering infection transmission chains. We systematically evaluated three widely adopted SARS-CoV-2 whole-genome sequencing methods for their sensitivity, specificity, and ability to reliably detect low-frequency variants. Our analyses reveal that the ARTIC v3 protocol consistently displays high sensitivity for generating complete genomes at low viral loads compared with the probe-based Illumina Respiratory Viral Oligo panel and a pooled long-amplicon method. We show substantial variability in the number and location of low-frequency variants detected using the three methods, highlighting the importance of selecting appropriate methods to obtain high-quality sequence data from low-viral-load samples for public health and genomic surveillance purposes.