a, Model diagram and duration of disease states in days, where d parameters represent the duration of time in each disease state (see Methods), yi is the fraction of infections that manifest as clinical cases in age group i, λi is the force of infection in age group i, PI is the incubation period and PS is the serial interval (see Methods). b, Susceptibility by age for the three models, with mean (lines), 50% (darker shading) and 95% (lighter shading) credible intervals shown. Age-specific values were estimated for model 1 (orange). Susceptibility is defined as the probability of infection on contact with an infectious person. c, Clinical fraction (yi) by age for the three models. Age-specific values were estimated for model 2 (blue) and fixed at 0.5 for models 1 and 3. d, Fitted contact multipliers for holiday (qH) and restricted periods (qL) for each model showed an increase in non-school contacts beginning on 12 January (start of the Lunar New Year) and a decrease in contacts following restrictions on 23 January. e, Estimated R0 values for each model. The red barplot shows the inferred window of spillover of infection. f, Incident reported cases (black) and modeled incidence of reported clinical cases for the three models fitted to cases reported by China Centers for Disease Control (CCDC)1 with onset on or before 1 February 2020. Lines mark the mean and the shaded window is the 95% highest density interval (HDI). g, Age distribution of cases by onset date as fitted to the age distributions reported by Li et al.27 (first three panels) and CCDC1 (fourth panel). Data are shown in open bars and model predictions in filled bars, where the dot marks the mean posterior estimate. h, Implied distribution of subclinical cases by age for each model. Credible intervals on modeled values show the 95% HDIs; credible intervals on data for g and h show 95% HDIs for the proportion of cases in each age group.