Fig. 4

Persistence of Delta variants in public sequence data after the emergence of Omicron. A) Temporal dynamics of VOCs during the course of the pandemic based on 13,342,630 high-quality SARS-CoV-2 whole genome sequences from the GISAID database collected between December 2019 and February 2023. B) Temporal dynamics of non-Omicron SARS-CoV-2 genome sequences collected after June 2022. C) Maximum likelihood phylogenetic tree of non-Omicron variants collected after June 2022 (triangles; n = 226) and random subsamples of the same variants collected around their respective peak prevalence (circles; n = 2,195). D) Lineages-Through-Time (LTT) plot derived from Bayesian analysis of SARS-CoV-2 Delta variant genome sequences collected around the peak of prevalence (n = 200) and after June 2022 (n = 200). The plot chronologically represents the accumulation of lineages from September 2020 to March 2023. E) Violin plots displaying pairwise p-distance distributions of SARS-CoV-2 Delta variant genome sequences. Distributions represent all pairwise comparisons between genomes from the peak of prevalence (n = 200), post-June 2022 (n = 200), and the pairwise divergence between these two time points. Each “violin” width is indicative of the kernel density estimation of the data, presenting an overview of genetic divergence within and between the two sample groups. Thick line in each violin shows median pairwise p-distance of each group and whiskers indicate the upper and lower 1.5 interquartile ranges (IQR). P-values were calculated using the Wilcoxon rank-sum test and subjected to adjustment for False Discovery Rate (FDR) using the Benjamini-Hochberg procedure, accounting for multiple comparisons