Plagued by a cryptic clock: insight and issues from the global phylogeny of Yersinia pestis

Abstract

Plague has an enigmatic history as a zoonotic pathogen. This infectious disease will unexpectedly appear in human populations and disappear just as suddenly. As a result, a long-standing line of inquiry has been to estimate when and where plague appeared in the past. However, there have been significant disparities between phylogenetic studies of the causative bacterium, Yersinia pestis, regarding the timing and geographic origins of its reemergence. Here, we curate and contextualize an updated phylogeny of Y. pestis using 601 genome sequences sampled globally. Through a detailed Bayesian evaluation of temporal signal in subsets of these data we demonstrate that a Y. pestis-wide molecular clock is unstable. To resolve this, we developed a new approach in which each Y. pestis population was assessed independently, enabling us to recover substantial temporal signal in five populations, including the ancient pandemic lineages which we now estimate may have emerged decades, or even centuries, before a pandemic was historically documented from European sources. Despite this methodological advancement, we only obtain robust divergence dates from populations sampled over a period of at least 90 years, indicating that genetic evidence alone is insufficient for accurately reconstructing the timing and spread of short-term plague epidemics.

Fig. 1. Global phylogeny of Yersinia pestis.

The phylogenetic and spatiotemporal diversity of 601 Y. pestis genomes. Populations were defined by integrating three nomenclature systems: the major branches, biovars, and time periods. A The maximum likelihood phylogeny of Y. pestis with strict bifurcations (no polytomies) and branch lengths scaled by genetic distance from the root in the number of nucleotide substitutions per site. The tree was rooted using two genomes of the outgroup taxa Y. pseudotuberculosis, which were pruned before visualization. B The mean sampling age of each genome with internal node dates bounded by ancient DNA calibrations. C The sampling location of each genome with coordinates standardized to the centroid of the associated province/state. The map image is copyright Mapbox and Open Street Maps used with permission. All images are produced by nextstrain.org are by CC BY.

Fig. 2. Substitution rate variation in Y. pestis.

A A root-to-tip regression on mean sampling age using all genomes from the maximum likelihood phylogeny. B A root-to-tip regression on mean sampling age by population. The distance to the population MRCA was calculated using subtrees extracted from the maximum likelihood phylogeny. C Bayesian substitution rates within and between populations. For each branch in the maximum clade credibility (MCC) trees, we extracted the estimated substitution rate (subs/site/year) and converted this to subs/year based on an alignment of 4,229,098 genomic sites.

Fig. 3. Ancestor-descendant relationships in the maximum likelihood phylogeny reveal tMRCA conflicts between Antiqua (0.ANT) and the First Pandemic (0.ANT4).

Node dates (95% HPD) were estimated from the Bayesian analysis, where each population was assessed independently. Grey branches indicate outliers, as defined by the 90% confidence interval of external branch lengths from all populations.

Fig. 4. The geographic distributions of the Second Pandemic; (1.PRE), Medievalis (2.MED), and Pestoides (0.PE) populations.

The sampling location of each genome was standardized to the centroid of the associated province and/or state. The map image is copyright Mapbox and Open Street Maps used with permission. All images are produced by nextstrain.org are by CC BY.

Fig. 5. Association between host and country of origin.

Linear regression of host associations. A linear regression of host-associations on the degree of geographic structure at the country level.

Fig. 6. Geographic origins and spread of the Third Pandemic (1.ORI) and the intermedium (1.IN) population.

Ancestral locations were estimated by fitting a discrete migration model to the maximum likelihood phylogeny using sampling locations by province. Arrows reflect the directionality of spread, but not the precise route taken. High-confidence migrations estimated with an ancestral likelihood greater than or equal to 0.95 and a branch support bootstrap of greater than or equal to 95% are shown. copyright Mapbox and Open Street Maps used with permission. All images are produced by nextstrain.org are by CC BY.

Fig. 7. Geographic origins and spread of the Second Pandemic (1.PRE), the ancestral Antiqua (0.ANT) and descendant Intermedium (1.IN) populations.

Ancestral locations were estimated by fitting a discrete mugration model to the maximum likelihood phylogeny using sampling locations by province. Arrows reflect the directionality of spread, but not the precise route taken, for high-confidence migrations estimated with an ancestral likelihood greater than or equal to 0.95 and a branch support bootstrap of greater than or equal to 95%.