Tracking the COVID-19 epidemic in Switzerland: phylogenetics and epidemiological modeling

Background:In Switzerland, SARS-CoV-2 was first detected in late February, and the COVID-19 epidemic has grown to 10,456 confirmed cases and 145 reported deaths as of 25 March 2020. Across Europe, case numbers and fatalities have been doubling rapidly for several weeks. Alongside the all important patient care by health care professionals, scientists from all over the world have developed mathematical models of COVID-19 transmission and sequenced numerous viral genomes to track the spread and evolution of SARS-CoV-2. During the ongoing COVID-19 pandemic, Nextstrain has developed into the primary platform for real-time analysis of all available SARS-CoV-2 genomes. Sequence submissions from all over the world now paint a detailed phylogeographic picture and clearly show how outbreaks in different parts of the world are related. At the same time, epidemiological modeling efforts have provided essential input for policy makers and anticipated the scale of the outbreak early on. While these two computational approaches have the potential to inform each other, these efforts have been largely disjoint to date.Aims:We aim to integrate genomic epidemiology and epidemiological modeling in order to provide comprehensive inferences and projections for the COVID-19 epidemic in real-time. Specifically, we will apply genomic epidemiology to conduct phylogenetic analysis of SARS-CoV-2 sequence data specifically for Switzerland, apply epidemiological modeling to provide real-time analysis and projections of the COVID-19 epidemic in Switzerland, and expand the Nextstrain platform by an integrated modeling approach where phylogenetic analysis and epidemiological modeling inform each other.Methods:The proposal builds on well-established methodologies from two research teams that have made important contributions to the epidemiological understanding of the COVID-19 pandemic. Phylogenetic techniques are well established in the analysis of viral genomes and Nextstrain has shown how such analyses can be done in real-time. Detailed epidemiological models that are embedded in maximum likelihood and Bayesian frameworks offer great potential for real-time analyses and short-term projections of the spread of emerging infectious diseases.Impact:With continuing spread of SARS-CoV-2 and its consequences for people's health, the economy and the society, it will be critically important that decision makers have a detailed picture of the ongoing epidemiological situation during different phases of the epidemic. The integrated solution for real-time analysis that combines genomic epidemiology with epidemiological modeling within the Nextstrain platform will provide actionable information for decision makers, public health authorities, and for hospital planning. We will be able to track the COVID-19 epidemic across age groups and cantons, and identify transmission chains and links to outbreaks outside of Switzerland.The direct integration with epidemiological models will allow quantification of the phylogenetic information in terms of numbers of infected individuals, or stochasticity of transmission at high spatio-temporal resolution. Such fine-grained analyses and projections will be particularly important when infection control moves from generalized interventions to targeted measures. Finally, the resulting infrastructure will be available for the surveillance and management of other pathogens in the future.