Primary author: Lily Geidelberg

Olivia Boyd, Lily Geidelberg, David Jorgensen, Manon Ragonnet, Igor Siveroni, Erik Volz and the Imperial College COVID-19 Response Team

Report prepared on 2020-04-28

Key points

  • Based on the genetic diversity of 47 whole SARS-CoV-2 genomes sampled in Wisconsin up to April 6th, we estimated several key epidemiological parameters
  • On April 6th, we estimate the total cumulative infections in Wisconsin to be 20495 [7623-86820], compared to the official reported total of 2440, reflecting a reporting rate of around 10%
  • From early February until early March, we estimate a reproduction number above 2, which declines to 1.35 [0.451-1.98] by April 6th
  • These results are preliminary and more reliable estimates will be produced when more cases in Wisconsin have been sequenced

Background information

This is analysis is based on :

  • 47 whole genomes sampled from within Wisconsin
  • 75 whole genomes sampled from outside of Wisconsin
  • Samples within Wisconsin were collected between 2020-03-14 and 2020-04-06

The sequences were extracted from GISAID on the 18th of April 2020; since then, there will have been further sequences uploaded to Wisconsin. Further, certain sequences downloaded were excluded from the analysis as we removed deuplicates, those with likely sequencing errors, or significant gaps.

The figure below shows the dates of sampling of the Wisconsin and external sequences.

plot of chunk sampling dist

Figure 1: Sampling distributions over time of number of sequences included within the region versus sequences included from the international reservoir.

How many are infected in Wisconsin?

Fitting a phylodynamic model to Wisconsin and international SARS-CoV-2 sequence data, we estimate epidemiological parameters. In Wisconsin, at the date of the last sample (2020-04-06), We estimate the total cumulative infections to be 20495 [7623-86820] median [95%CI]; the corresponding official reported total on the same day was 2440.

plot of chunk Cumulative estimated infections through time log scale

Figure 2: Estimated cumulative infections through time represented by solid black line (median) and 95% CrI (ribbon). Black points represent reported cases in Wisconsin. The dashed line indicates the date of last sample in Wisconsin in this analysis.

plot of chunk daily estimated infections through time log scale

Figure 3: Estimated daily infections through time represented by solid black line (median) and 95% CrI (ribbon). Black points represent reported cases in Wisconsin. The dashed line indicates the date of last sample in Wisconsin in this analysis.

We estimate the cumulative and daily number of infections about an order of magnitude above the official reported statistics. This underreporting is likely due to asymptomatic infections who do not present to hospital. The figure below shows the estimated reporting rate, which compares the reported and estimated number of infections in Wisconsin.

plot of chunk reporting

Figure 4: Estimated percentage of daily cases reported in Wisconsin. Error bars represent the 95% credible interval.

To understand the rate of spread of SARS-CoV-2, we estimated the reproduction number over time (Rt). Our estimate including credible intervals for Rt was above 2 until early March, when it starts to decrease. By the last sample, we estimate Rt to be 1.35 [0.451-1.98] median [95% CrI].

plot of chunk Rt

Figure 5: Reproduction number through time. The black vertical dashed line indicates the date of last sample in Wisconsin in this analysis. Orange and red dashed lines indicate dates of school closure and general lockdown in Wisconsin, respectively.

How quickly has the epidemic in Wisconsin grown?

Quantile Reproduction number Growth rate (per day) Doubling time (days)
50% 2.29 0.130 5.34
2.5% 1.96 0.101 4.42
97.5% 2.62 0.157 6.89

Table 1: Reproduction number, growth rate and doubling times

How has SARS-CoV 2 evolved in Wisconsin?

The figure below represents the time scaled phylogeny that is co-estimated in our analysis. We observe that samples in Wisconsin (red points) feature in several places in the phylogeny, which indicates importations from outside the state.

plot of chunk mcc_tree

Figure 6: Time scaled phylogeny co-estimated with epidemiological parameters. The colour of the tips corresponds to location sampling; red tips were sampled from within Wisconsin, blue tips from outside.

Molecular clock rate of evolution: 0.00123 [0.000982-0.00152] median [95% CrI]

Methods summary

Details on methods and priors can be found here.

Statistic mean ESS
posterior -43170 131
likelihood -43056 993
prior -114.2 103
treeLikelihood -43056 993
TreeHeight 0.2906 97
clockRate 0.001232 108
kappa 5.512 18353
PhydynSEIR -82.37 121
seir.E 5.045 101
seir.S 91038 204
seir.b 17.47 76
seir.exog 0.006643 72
seir.exogGrowthRate 26.22 39
seir.importRate 8.803 1789
seir.p_h 0.2094 86
seir.tau 74.07 144
freqParameter.1 0.2976 5551
freqParameter.2 0.1824 6024
freqParameter.3 0.1952 6393
freqParameter.4 0.3248 5161
gamma0 73 NA
gamma1 121.7 31

Table 2: Effective sample size of model parameters

Model version: seijr0.1.0

Report version: 20200428-120357-6c31f383


This work was supported by the MRC Centre for Global Infectious Disease Analysis at Imperial College London.

Sequence data were provided by GISAID and these laboratories.