Timing the SARS-CoV-2 Index Case in Hubei Province

Abstract

Understanding when SARS-CoV-2 emerged is critical to evaluating our current approach to monitoring novel zoonotic pathogens and understanding the failure of early containment and mitigation efforts for COVID-19. We employed a coalescent framework to combine retrospective molecular clock inference with forward epidemiological simulations to determine how long SARS-CoV-2 could have circulated prior to the time of the most recent common ancestor. Our results define the period between mid-October and mid-November 2019 as the plausible interval when the first case of SARS-CoV-2 emerged in Hubei province. By characterizing the likely dynamics of the virus before it was discovered, we show that over two-thirds of SARS-CoV-2-like zoonotic events would be self-limited, dying out without igniting a pandemic. Our findings highlight the shortcomings of zoonosis surveillance approaches for detecting highly contagious pathogens with moderate mortality rates.

Fig. 1.

Hypothetical coalescent scenarios depicting how the time between index case infection and time of stable coalescence can vary based on stochastic extinction events of basal viral lineages. Coalescence can occur within the index case (upper left) or in cases infected later in the course of the epidemic. In extreme cases, the epidemic can persist at low levels for a long time before coalescence (lower right).

Fig. 2.

Posterior distribution for the time of the most recent common ancestor (tMRCA) of 583 sampled SARS-CoV-2 genomes circulating in China between December 2019 and April 2020. Shaded area denotes 95% HPD. Long-dashed line is 17 November 2019, and short-dashed line is 01 December 2019.

Fig. 3.

Forward simulations estimating the timing of the index case in Hubei province. (A) Days between index case infection and stable coalescence in forward compartmental epidemic simulations (n=1000). (B) Days between index case infection and stable coalescence after rejection sampling, conditioned on an ascertained case by 17 November 2019. (C) Posterior distribution for date of index case infection, conditioned on an ascertained case by 17 November 2019 denoted by long-dashed line. (D) Days between index case infection and stable coalescence after rejection sampling, conditioned on an ascertained case by 01 December 2019. (E) Posterior distribution for date of index case infection, conditioned on an ascertained case by 01 December 2019 denoted by short-dashed line. Grey dashed lines indicate median estimates.

Fig. 4.

Combined simulation and phylogenetic workflows to estimate the timing of the Hubei index case. (1a) Using sequence and epidemiological data, (1b) BEAST performs a phylodynamic molecular clock analysis to (1c) determine the tMRCA. (2a) FAVITES simulates the epidemic in Hubei using a SAPHIRE compartmental model (22) and (2b) estimates a prior distribution for the time from index case to the stable coalescence. The results of (1) and (2) are combined via rejection sampling (3; Fig. S5) to (4) determine the timing of the index case and its posterior distribution.

Fig. 5.

Epidemic growth in compartmental simulations. (A) Total estimated people infected in late-2019. Dark purple shading is central 50% HPD, intermediate purple shading is central 95% HPD, and light purple is central 99% HPD. (B) The number of people infected over time in a sample of epidemic simulations that established (purple; n=30) and failed to establish (grey; n=70). The y-axis transitions to log-scale once ≥10 people are infected at any given time. The lower panel shows the proportion of simulations that still have ≥1 infected individual over time (persisting epidemics in purple; extinct epidemics in grey).