New Routes to Phylogeography: A Bayesian Structured Coalescent Approximation

Abstract

Phylogeographic methods aim to infer migration trends and the history of sampled lineages from genetic data. Applications of phylogeography are broad, and in the context of pathogens include the reconstruction of transmission histories and the origin and emergence of outbreaks. Phylogeographic inference based on bottom-up population genetics models is computationally expensive, and as a result faster alternatives based on the evolution of discrete traits have become popular. In this paper, we show that inference of migration rates and root locations based on discrete trait models is extremely unreliable and sensitive to biased sampling. To address this problem, we introduce BASTA (BAyesian STructured coalescent Approximation), a new approach implemented in BEAST2 that combines the accuracy of methods based on the structured coalescent with the computational efficiency required to handle more than just few populations. We illustrate the potentially severe implications of poor model choice for phylogeographic analyses by investigating the zoonotic transmission of Ebola virus. Whereas the structured coalescent analysis correctly infers that successive human Ebola outbreaks have been seeded by a large unsampled non-human reservoir population, the discrete trait analysis implausibly concludes that undetected human-to-human transmission has allowed the virus to persist over the past four decades. As genomics takes on an increasingly prominent role informing the control and prevention of infectious diseases, it will be vital that phylogeographic inference provides robust insights into transmission history.

Establishment and cryptic transmission of Zika virus in Brazil and the Americas

“…We found no significant differences in evolutionary rates among ZIKV genome regions (Extended Data Table 3a). The estimated ratio of divergence at nonsynonymous and synonymous sites (dN/dS) of the Am-ZIKV lineage is low (0.11, 95% confidence interval 0.10–0.13), as observed for other vector-borne flaviviruses 28 , but is higher than that of PreAm-ZIKV viruses (0.061, 0.047–0.077), probably owing to the raised probability of observing slightly deleterious changes in short-term datasets, as observed during previous epidemics 29 .…”

“…Phylogeographic reconstructions were conducted using two approaches; (i) using the asymmetric 52 discrete trait evolution models implemented in BEASTv1.8.4 46 and (ii) using the Bayesian structured coalescent approximation (BASTA) 29 implemented in BEAST2v.2. The latter has been suggested to be less sensitive to sampling biases 53 .…”

“…See Methods for details. ANLPPs were calculated using two approaches: DTA=discrete trait analysis method 30 (left side columns) and BASTA=Bayesian structured coalescent approximation method 29 (right side columns). For each method, we employed an asymmetric model of location exchange to estimate ancestral node locations and to infer patterns of virus spread among regions.…”

Effects of the population pedigree on genetic signatures of historical demographic events

“…The reasons for this are that gene genealogies are in fact the results of random processes, likely at the population level but certainly at the level of Mendelian genetic transmission, and that it is not known a priori whether a given set of data comes from one of those cases in which gene genealogies are individually informative (8-10). Although this particular issue may be considered settled, debates about the proper application of random models in phylogeography continue to arise (11,12).We consider an additional question about the application of random models that has received comparatively little attention either in phylogeography or population genetics. Namely, what is the extent to which genealogies in the family sense-also known as organismal pedigrees (13) or population pedigrees (14)-constrain gene genealogies and thus genetic variation?…”

“…The reasons for this are that gene genealogies are in fact the results of random processes, likely at the population level but certainly at the level of Mendelian genetic transmission, and that it is not known a priori whether a given set of data comes from one of those cases in which gene genealogies are individually informative (8-10). Although this particular issue may be considered settled, debates about the proper application of random models in phylogeography continue to arise (11,12).…”