Extended Data Fig. 1: Deviation from linear distribution of vaccine doses can (sometimes) affect evolution of resistance.
From: Evolution of resistance to COVID-19 vaccination with dynamic social distancing
(A) Shape of the vaccination curve in each of the six countries (Brazil, France, Germany, Israel, United Kingdom and United States) and straight line representing constant vaccination at rate c. We compare these 7 possibilities in panels B, C and D for 3 different c values. (B,C,D) The increase in the number of vaccinated individuals affects the increase of the reproductive ratio of the vaccine resistant mutant RMT, and hence the probability of emergence of a vaccine resistant mutant. See Supplementary Fig. 15 for further information how the vaccination rate affects the probability of vaccine resistance at a fixed time t. For each (L, c) parameter set and each of the six countries, we report on the probability of emergence of vaccine resistance, calculated as the proportion of runs, out of a 1000, where we observed the emergence of the MT strain, in the first 210 days, the approximate number of days since mass vaccination started. We observe no significant difference between linear vaccination (in gray in panels B, C and D) and any of the real vaccination curves from the six countries, scaled to the same value of c. We observe a significant difference in probability of emergence of vaccine resistance when considering data from Israel and when considering data from France, Germany and Brazil. The probability of emergence of vaccine resistance is higher for simulation run with Israeli dynamics of vaccine distribution for c = 5000 (see panel C) and lower for c = 10000 (see panel D). As shown in panel A, the number of vaccinated increased very sharply in a linear fashion in Israel, then plateaued at a high value. Hence, for high average number of new vaccinations per day c, the probability of emergence of vaccine resistance decreases compared to that observed with vaccine distribution dynamics from other countries. Parameters: N = 106; a = 0.25; d = 0.01; μ = 10−6; q = 1; s0 = 0.1, β1 = β2 = 7.5 ∙ 10−7. Error bars are calculated as p ± 1.96 ∙ √(p(1 − p)/1000), where p is the proportion of runs out of n=1000 where we observed emergence of the vaccine resistant mutant.
