European SARS-CoV-2 and influenza Bioinformatics External Quality Assessment (ESIB-EQA), 2023

SARS-CoV-2 and influenza viruses are pathogens of major public health impact. Surveillance has played a vital role in monitoring the COVID-19 pandemic as well as seasonal influenza, including public health actions such as updates on vaccines and antiviral drugs.

The objective of this External Quality Assessment (EQA) is to strengthen the capacity for genomic epidemiology and public health bioinformatics, which is crucial for response during a pandemic or unexpected major public health event. 

The EQA was divided into SARS-CoV-2 and influenza and was composed of different components that could be performed individually, together representing both short read and long read sequencing technology, and included the bioinformatics processes relevant for public health activities, i.e. consensus sequence generation and quality control, clustering and classification of genomes and mutation analysis for prediction of phenotypic properties. Laboratories from the 27 European Union (EU) and the three European Economic Area (EEA) countries as well as the six Western Balkan countries and Türkiye were invited to participate. The EQA was open between 1 March and 1 April 2023, and the correctness of the responses and throughput time was assessed.

The EQA had participation from 25 of 30 EU/EEA countries and two of six Western Balkan countries. The components they participated in varied. The quartile of the adjusted throughput time for these four components, indicative of the more performant laboratories, was in the range of one to two working days. A large majority of
the laboratories correctly classified the samples according to their expected quality status. For the consensus generation of SARS-CoV-2 genomes, issues were mostly related to insertions or deletions (‘indels’) in comparison to the comparator sequence and for influenza, the issues were mostly related to that consensus sequences were not generated in their full length and primer sequences were not considered artificial sequences. For the genome clustering component, a process important for outbreak investigations and for following virus evolution, laboratories performed well. For both SARS-CoV-2 and influenza, the most common discrepancies were due to using different thresholds of cluster definitions (number of mutations). The mutation analysis component for SARSCoV- 2 was in general accurate, however for influenza there was significant fraction of mutation-based misclassifications of resistance to antiviral drugs oseltamivir, zanamivir and baloxavir marboxil. 

The issues identified in this EQA may facilitate several improvements in the laboratories’ bioinformatic processes. The example of indels that are either falsely assigned or missed may lead to incorrect variant classification or delayed detection of an emerging variant or that cases are not assigned to an outbreak which they belong to. Furthermore, for influenza, standardisation of parameters such as flanking regions, primer sequence masking and antiviral susceptibility predictions can facilitate inter-laboratory comparisons. Novel viruses provided as educational samples that does not
affect the score may improve detecting reassortment and determination of subtype and origin. 

This EQA establishes a baseline that could be used for future comparisons. The participants received individual feedback with suggestions for improvement and laboratories from three countries were invited to and completed a five-day twinning training activity. Furthermore, general recommendations are compiled in this report in order to improve the laboratory practices and performance.