Fig. 3: For the Latin America cohort, vaccine efficacy (VE) against the primary COVID-19 endpoint by different physicochemical-weighted (PCW) Hamming distances.

The Hamming distances were from the disease-causing SARS-CoV-2 isolate to that of the vaccine-insert sequence, for the following regions: a Spike, b the RBD domain, c the NTD domain, or d the S1 region. The top plot in each panel shows the distributions of distances by treatment arm, color-coded by lineage: Alpha, yellow; Epsilon, brown; Gamma, turquoise; Lambda, red; Mu, green; Reference, gray; Zeta, blue. The left and right edges of the box plots represent the 25th and 75th percentiles of PCW-Hamming distance in the designated region, and the vertical middle line represents the 50th percentile. The horizontal bars extend from the 25th (or 75th) percentile of PCW-Hamming distance to the minimum (or maximum) PCW-Hamming distance within the 25th (or 75th) percentile of Hamming distance minus (or plus) 1.5 times the interquartile range. The bottom plot in each panel shows the estimated VE by SARS-CoV-2 sequence distance. The dotted lines are pointwise 95% confidence intervals. The dots are overall VE estimates for the given lineage placed at the lineage-specific median distance of placebo arm endpoints, with vertical bars indicating their pointwise 95% confidence intervals. In each panel, the “Double 1-sided unadjusted sieve p value” doubles the p value from a one-sided Wald test of the null hypothesis of constant VE vs. the alternative hypothesis of a decreasing VE with an increasing value of the feature on the x-axis (Juraska and Gilbert¹⁹, Section 5). Two Zeta sequences are visible outliers from other Zeta sequences; both sequences have two large deletions (9AA and 7AA in length) in the N-terminal domain. The plots reveal that Lambda has two sub-lineages, one (n = 79) with a range of distances 17.2–18.9 and a second (n = 9) with a range of distances 25.8–27.7, due to a 13-AA deletion between sites 64 and 76. CI confidence interval, FWER family-wise error rate, NTD N-terminal domain, RBD receptor-binding domain, Unadj. unadjusted.