Table 3.
Values and priors for the parameters of the BDSKY model. s/s/y=substitutions per site per year
| Parameter | Dataset | Value or prior | Rationale/Assumption |
|---|---|---|---|
| Clock rate | Brazil | 4.0 × 10−4 s/s/y | Informed by root-to-tip regression |
| Hong Kong | 1.0 × 10−4 s/s/y | ||
| Death rate | Brazil and Hong Kong | 36.5 y−1 | The period between infection and becoming uninfectious assumed an exponential distribution with a mean of 10 days15 |
| Reproduction number | Brazil and Hong Kong | Lognormal (0.8, 0.5) | Median 2.2, 95% IQR 0.8–5.9 |
| Time of origin | Brazil | Lognormal (−1.50, 0.4) y before present | Median 4 December 2020, 95% IQR 25 September 2020 to 12 January 2021 |
| Hong Kong | Lognormal (−1.75, 0.4) y before present | Median 18 January 2020, 95% IQR 17 November 2019 to 15 February 2020 | |
| Sampling proportion | Brazil | Uniform (0, 0.024) | 196 sequences from 8246 suspected P.1 cases as of 7 February 2021 |
| Hong Kong | Uniform (0, 0.116) | 117 sequences from 1012 confirmed cases as of 7 May 2020 |