Fig. 3: Dynamic immune landscape and variant dynamics in Germany.
a, Historical variant dynamics in Germany during 1 March 2022 to 1 July 2024 of spike pseudo-groups BA.2,.X BA.5.X (+ BA.4.X + BE.1.1), BF.7.X, BQ.1.1.X, XBB.1.5.X (+ EG.1.X), XBB.1.9.X, EG.5.X, JN.1.X (+ BA.2.86.X) and KP.X, where ‘.X’ denotes inclusion of the respective sublineages. Confidence intervals were computed using Wilson’s method. Insets show model-predicted relative fitness γy(t) (representing the relative number of susceptible individuals; coloured areas), with superimposed changes in frequency (πy(t − 1)/πy(t) − 1) seen in the data (solid lines). Note that the sequencing intensity changed markedly in April 2023 (vertical dashed line) as highlighted in Extended Data Fig. 3. b, Predicted variant dynamics. Left: model-predicted relative growth advantage γy of emerging spike pseudo-groups XBB.1.5* (+ EG.1*), XBB.1.9.X (+ EG.1.3), XBB.1.16.X, EG.5.1* and BA.2.86 (colours in key) using data until 15 April 2023 (asterisks (for example, EG.1*) indicate spike pseudo groups; that is, the group of lineages with identical mutations in the spike protein). Note that BA.2.86 was first detected on 24 July 2023 in Denmark and represents the dominating lineage (together with daughter lineage JN.1 globally, as of January 2024). Thick lines show average values, and light lines show minimum–maximum ranges resulting from pharmacokinetic variability. Right: data-derived lineage frequencies in the time after the prediction horizon (15 April 2023 to 27 July 2023). Confidence intervals (95%) were calculated using Wilson’s method with sample size n = 2,919, 500, 165 and 164 for April, May, June and July, respectively.
