Challenges in estimating virus divergence times in short epidemic timescales with special reference to the evolution of SARS-CoV-2 pandemic.

Carlos G Schrago,Lucia P Barzilai

doi:10.1590/1678-4685-gmb-2020-0254

Carlos G Schrago, Lucia P Barzilai

Open Access

https://doi.org/10.1590/1678-4685-gmb-2020-0254

Copy DOI

Abstract

The estimation of evolutionary parameters provides essential information for designing public health policies. In short time intervals, however, nucleotide substitutions are ineffective to record all complexities of virus population dynamics. In this sense, the current SARS-CoV-2 pandemic poses a challenge for evolutionary analysis. We used computer simulation to evolve populations in scenarios of varying temporal intervals to evaluate the impact of the age of an epidemic on estimates of time and geography. Before estimating virus timescales, the shape of tree topologies can be used as a proxy to assess the effectiveness of the virus phylogeny in providing accurate estimates of evolutionary parameters. In short timescales, estimates have larger uncertainty. We compared the predictions from simulations with empirical data. The tree shape of SARS-CoV-2 was closer to shorter timescales scenarios, which yielded parametric estimates with larger uncertainty, suggesting that estimates from these datasets should be evaluated cautiously. To increase the accuracy of the estimates of virus transmission times between populations, the uncertainties associated with the age estimates of both the crown and stem nodes should be communicated. We place the age of the common ancestor of the current SARS-CoV-2 pandemic in late September 2019, corroborating an earlier emergence of the virus.

Highlights

The evolutionary analysis of virus genomes frequently relies on molecular phylogenies, which illustrate the ancestry of lineages in tree graphs (Holmes, 2008)
We investigated the extent to which the age of the of the virus epidemic affects the inference of evolutionary parameters, in order to elucidate whether the discrepancies between estimates of SARS-CoV-2 timescales may be caused by the stochasticity of coalescent process and the reduced genetic diversity in narrow timescales
Our simulation showed that the timescale of the epidemic significantly impacted the estimates of evolutionary parameters of epidemiological interest (Figure 2)

Summary

Introduction

The evolutionary analysis of virus genomes frequently relies on molecular phylogenies, which illustrate the ancestry of lineages in tree graphs (Holmes, 2008). A time direction, implying ancestor-to-descendent relationship, is incorporated into phylogenies. In order to fully incorporate the temporal dimension onto trees, divergence times of nodes must be estimated. By employing some calibration information, a direct linear transformation may be readily applied. This is the standard molecular clock, in which sequence substitution rates are constant along branches and across lineages. Timescales may be inferred by accommodating rate variation among lineages (Gillespie, 1991; Kishino et al, 2001; Bromham et al, 2018)

Methods

Results

Conclusion