Tempo and Trajectories of SARS-CoV-2 Evolution: A Six-Year Genomic Analysis (2019–2025) of 1.13 Million Genomes

Background: Even after six years of SARS-CoV-2 circulation and successive Variants of Concern (VOCs)—including Alpha, Delta, and Omicron lineages (BA.*, XBB, JN.1, XFG)—the coronavirus continues to challenge public health. Large-scale analysis is essential for understanding adaptive mechanisms and the underlying mutational processes. Objectives: To address this, we quantify mutation frequencies at each genomic site and track how these mutations change over time across the pandemic. Using emergence slopes, we identify mutations that emerge rapidly and investigate how the overall spectrum of mutations has shifted throughout the pandemic. Methods: We analyzed 1.13 million complete SARS-CoV-2 genomes (NCBI Virus; December 2019–September 2025) with monthly binning. Reference-based alignment (NC_045512.2) was performed using Minimap2 and SAMtools at uncapped pileup depth. Emergence slopes were calculated by linear regression of mutation frequency on time, and alleles rising over the most recent twelve months were modelled individually by quasi-binomial logistic regression. Results: The genome-wide landscape revealed hotspot mutations in Spike (S) and Nucleocapsid (N) proteins. Distinct evolutionary phases corresponding to VOC dominance were also identified. The mutation spectrum was consistently dominated by C>T substitutions, while changes in G>T substitution rates were associated with lineage turnover, including a marked decrease following the Delta period. The emergence-slope landscape highlights concentrated adaptive signals in Spike. Because frequency is bounded, the overall slope has an arithmetic maximum attained by a sweep centered on the middle of the observation window; the highest slope sites are Spike and ORF1a positions that exceeded 99% frequency between May 2022 and December 2023 and have remained there since. Over the most recent twelve months, 14 alleles rose from below 5% to above 50%. Eight are synonymous and all belong to a single recombinant (XFG) haplotype, which the allele set recovers with 97.7–100% sensitivity against lineage assignment. Conclusions: SARS-CoV-2 continues evolving quickly six years after its appearance, with the S protein remaining the key target for adaptation. Variations in the spectrum show how the virus–host interaction occurs in real time (such as shifts in replication niche). Allele-frequency metrics applied over a recent window provide scalable, phylogeny-independent monitoring that recovers lineage definitions without a tree.