Reconstructing relative transmission rates in Bayesian phylodynamics: Two-fold transmission advantage of Omicron in Berlin, Germany during December 2021

Abstract

Abstract Phylodynamic methods have lately played a key role in understanding the spread of infectious diseases. During the coronavirus disease (COVID-19) pandemic, large scale genomic surveillance has further increased the potential of dynamic inference from viral genomes. With the continual emergence of novel severe acute respiratory syndrome coronavirus type 2 (SARS-CoV-2) variants, explicitly allowing transmission rate differences between simultaneously circulating variants in phylodynamic inference is crucial. In this study, we present and empirically validate an extension to the BEAST2 package birth–death skyline model (BDSKY), BDSKY$\lambda $, which introduces a scaling factor for the transmission rate between independent, jointly inferred trees. In an extensive simulation study, we show that BDSKY$\lambda $ robustly infers the relative transmission rates under different epidemic scenarios. Using publicly available genome data of SARS-CoV-2, we apply BDSKY$\lambda $ to quantify the transmission advantage of the Omicron over the Delta variant in Berlin, Germany. We find the overall transmission rate of Omicron to be scaled by a factor of two with pronounced variation between the individual clusters of each variant. These results quantify the transmission advantage of Omicron over the previously circulating Delta variant, in a crucial period of pre-established non-pharmaceutical interventions. By inferring variant- as well as cluster-specific transmission rate scaling factors, we show the differences in transmission dynamics for each variant. This highlights the importance of incorporating lineage-specific transmission differences in phylodynamic inference.