# Enhancing public health surveillance: a comparative study of platform-specific and hybrid assembly approaches in SARS-CoV-2 genome sequencing

## TL;DR

This study compares sequencing methods for SARS-CoV-2 to improve genome accuracy and public health responses.

## Contribution

The study systematically evaluates hybrid assembly for SARS-CoV-2 surveillance and identifies conditions for its optimal use.

## Key findings

- Hybrid assembly produced more complete genomes than Illumina-based methods.
- Hybrid assembly reliably detected mutations across all methodologies.
- High-quality ONT reads are crucial for successful hybrid assembly.

## Abstract

During the COVID-19 pandemic, next-generation sequencing (NGS) has been instrumental for public health laboratories in tracking severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) mutations and informing responses. Illumina systems and Oxford Nanopore Technologies (ONT) have been primary tools for NGS, each presenting unique advantages. The hybrid assembly (HA) approach, integrating short- and long-read sequencing methods, has been developed to improve genome accuracy by utilizing the combined advantages of both techniques. While HA has been used to enhance SARS-CoV-2 genome quality, its optimal applications for SARS-CoV-2 sequencing and surveillance have not been systematically studied. This study seeks to address this gap by evaluating the conditions under which HA improves SARS-CoV-2 genomic surveillance, analysing 192 samples using eight bioinformatics methods across both platforms. HA was evaluated against single-technology approaches for its genome assembly and mutation detection performance. While HA did not outperform single-technology methods in detecting unique mutations, it produced marginally more complete genomes than Illumina-based methods. Importantly, mutations identified by HA were consistently detected across all eight methodologies, demonstrating its reliability in mutation detection. Moreover, our research underlines the critical need for in-house validation of methods and exposes the limitations inherent in proprietary pipelines. Our findings suggest that an HA approach could be used as a quality control tool in genomic surveillance, particularly for improving low-quality ONT sequencing data by integrating high-quality Illumina sequencing data. However, implementing HA demands the presence of both sequencing platforms and additional resources, such as hands-on time, expensive sequencing reagents and bioinformatics know-how. A decision-tree analysis identified the percentage of trimmed ONT reads relative to total reads as crucial for HA success, emphasizing the significance of high-quality ONT reads. This comprehensive approach provides public health laboratories insights to refine genomic surveillance strategies for SARS-CoV-2, potentially influencing future research and response efforts.