Towards increased accuracy and reproducibility in SARS-CoV-2 next generation sequence analysis for public health surveillance

During the COVID-19 pandemic, SARS-CoV-2 surveillance efforts integrated genome sequencing of clinical samples to identify emergent viral variants and to support rapid experimental examination of genome-informed vaccine and therapeutic designs. Given the broad range of methods applied to generate new viral genomes, it is critical that consensus and variant calling tools yield consistent results across disparate pipelines. Here we examine the impact of sequencing technologies (Illumina and Oxford Nanopore) and 7 different downstream bioinformatic protocols on SARS-CoV-2 variant calling as part of the NIH Accelerating COVID-19 Therapeutic Interventions and Vaccines (ACTIV) Tracking Resistance and Coronavirus Evolution (TRACE) initiative, a public-private partnership established to address the COVID-19 outbreak. Our results indicate that bioinformatic workflows can yield consensus genomes with different single nucleotide polymorphisms, insertions, and/or deletions even when using the same raw sequence input datasets. We introduce the use of a specific suite of parameters and protocols that greatly improves the agreement among pipelines developed by diverse organizations. Such consistency among bioinformatic pipelines is fundamental to SARS-CoV-2 and future pathogen surveillance efforts. The application of analysis standards is necessary to more accurately document phylogenomic trends and support data-driven public health responses.

Competitive fitness of emerging SARS-CoV-2 variants is linked to their Distinctiveness relative to preceding lineages from that region

The COVID-19 pandemic has seen the persistent emergence of immune-evasive SARS-CoV-2 variants under the selection pressure of natural and vaccination-acquired immunity. However, it is currently challenging to quantify how immunologically distinct a new variant is compared to all the prior variants to which a population has been exposed. Here we define Distinctiveness of SARS-CoV-2 sequences based on a proteome-wide comparison with all prior sequences from the same geographical region. We observe a correlation between Distinctiveness relative to contemporary sequences and future change in prevalence of a newly circulating lineage (Pearson r = 0.75), suggesting that the Distinctiveness of emergent SARS-CoV-2 lineages is associated with their competitive fitness. We further show that the average Distinctiveness of sequences belonging to a lineage, relative to the Distinctiveness of other sequences that occur at the same place and time (n = 944 location/time data points), is predictive of future increases in prevalence (AUC = 0.88, 0.86-0.90 95% confidence interval). By assessing the Delta variant in India versus Brazil, we show that the same lineage can have different Distinctiveness-contributing positions in different geographical regions depending on the other variants that previously circulated in those regions. Finally, we find that positions that constitute epitopes contribute disproportionately (20-fold higher than the average position) to Distinctiveness. Overall, this study suggests that real-time assessment of new SARS-CoV-2 variants in the context of prior regional herd exposure via Distinctiveness can augment genomic surveillance efforts.