Genomic surveillance of SARS-CoV-2 variants using pooled WGS

Abstract

This study presents the development and validation of a genomic surveillance strategy using Whole Genome Sequencing (WGS) on normalized pooled samples to detect and monitor SARS-CoV-2 variants. A bioinformatics pipeline was designed specifically for analyzing pooled WGS data and was validated using simulated datasets, pooled samples of reference materials, and pooled clinical samples collected during key periods of the Delta and Omicron variant emergence. The approach was evaluated for its accuracy in estimating variant abundance at both the Phylogenetic Assignment of Named Global Outbreak (PANGO) lineage level and the World Health Organization (WHO) variant level. From the simulation datasets, the method achieved an overall sensitivity of 99.1% and a positive predictive value (PPV) of 99.9% for detecting SARS-CoV-2 variants at the WHO variant level. At the PANGO lineage level, it achieved an overall sensitivity of 82.8% and a PPV of 77.4% when a predicted lineage was considered accurate if it shared more than 90% of markers with any true lineage present in the pooled sample. The accuracy of variant abundance estimation was further validated using pooled samples of reference materials. Analysis of pooled clinical samples showed results consistent with national epidemiological trends, particularly during the emergence of the Delta and Omicron variants in Korea. This pooled WGS-based genomic surveillance strategy offers a scalable and economical solution for monitoring SARS-CoV-2 variants, providing public health authorities with a valuable tool for tracking pandemic dynamics and enabling timely responses.

Keywords: Epidemiological surveillance; SARS-CoV-2; SARS-CoV-2 mutation screening.

Fig. 1

Example of Result for COVID-19 Variant Abundance Estimation and Mutation Detection through Pooled NGS Data Analysis. This figure shows the outcomes of analyzing pooled NGS data. It provides the estimated abundance for each cluster of lineages and lists all detected mutations, each with its corresponding lineage associations.

Fig. 2

Lineage-Level Performance in COVID-19 Variant Detection. This figure shows the detection performance for key PANGO lineages associated with WHO COVID-19 variants, indicating both exact (superscript a) and approximate (superscript b) match outcomes for Detected Number, Sensitivity, and PPV (Positive Predictive Value). An approximate match is identified by a shared marker ratio exceeding 90%. The figure highlights the efficiency of detecting lineages such as Alpha (B.1.1.7), Delta (AY.4), and Omicron (BA.1) variants, demonstrating the system’s adaptability to both precise and nearly matching genetic sequences.

Fig. 3

Performance Evaluation of WHO Variant Level COVID-19 Variation Detection. This figure presents a comprehensive analysis of the sensitivity and positive predictive value (PPV) for the detection of various COVID-19 variants, including Alpha, Beta, Delta, Epsilon, Eta, GH/490R, Gamma, Iota, Kappa, Lambda, Mu, Omicron, and Zeta. The assessment was based on a total of 1120 tests, with the count of True Positives (TP), False Negatives (FN), and False Positives (FP) summarized for each variant. Sensitivity and PPV are depicted for each variant, highlighting the high accuracy and reliability of the detection method across the spectrum of COVID-19 variants, with most variants showing near-perfect sensitivity and PPV rates.

Fig. 4

Temporal Trends in WHO COVID-19 Variant Prevalence Through Simulated Pooled NGS Data. This figure shows estimated fractions of WHO-recognized COVID-19 variants from simulated pooled NGS, including Lambda, Mu, Omicron, GH/490R, Iota, Kappa, Epsilon, Eta, Gamma, Alpha, Beta, Delta, and Zeta. The simulation spans several months, illustrating the dynamic shifts in variant prevalence. Each variant’s estimated fraction is plotted over time, highlighting the changing landscape of COVID-19 variants.

Fig. 5

Reference Panel Evaluation for COVID-19 Variant Detection. This figure illustrates the performance of six reference panels in accurately estimating the fraction of WHO-designated COVID-19 variants, including Alpha, Beta, Delta, Gamma, and Omicron. For each panel, the figure compares real versus estimated values, showing the precision of detection across six different panels. It highlights the ability of the reference panels to closely estimate the actual fraction of variants.

Fig. 6

Analysis of Clinical Samples for COVID-19 Variant Prevalence. This figure demonstrates the prevalence of COVID-19 variants Omicron and Delta over a series of weeks, contrasting pooled clinical samples with official South Korean national prevalence data. It visually represents the proportion of each variant detected, emphasizing the shifting dominance of variants over the observe period. The side-by-side comparison with national statistics highlights the dynamic nature of the pandemic’s evolution within South Korea, reinforcing the essential role of ongoing surveillance in guiding public health strategies and interventions.