Phylogeography is a field of study concerned with the principles and processes governing the geographic distributions of genealogical lineages, especially those within or among closely related species. [Avise et al. Ann Rev Ecol Syst 1987]
Analyses of epidemic spread through space and time are often performed by defining a finite, non-ordered set of locations (e.g. countries) and using ancestral character reconstruction (ACR) [more details] along a phylogenetic tree with the defined locations as possible character states, based on the known tip locations. ACR aims to unravel how the character has changed on the tree from the root to the tips through time, by assigning the most likely ancestral character state(s) to each internal node.
Using our recently developed ancestral scenario reconstruction tool PastML [Ishikawa et al. MBE 2019, pastml.pasteur.fr], we study the phylogeography of various pathogens. In [Ishikawa et al. MBE 2019] we analysed the phylogeography of DENV2 epidemics and checked the robustness of ACR results regarding state sampling variations. In [Zhukova et al. 2019 (in press)] we performed an analysis of temporal and phylogeographic spread of all the four Dengue virus serotypes and compared the strengths and potential pitfalls of the two frequently-used approaches: maximum likelihood and Bayesian inference, highlighting state-of-the-art efforts to perform such computations more efficiently.
We collaborate with PANGEA consortium [Pillay et al. Lancet Infect Dis 2015, pangea-hiv.org] to study the phylogeography of HIV-1 in Africa. We also analyse the geographic spread of HIV-1 epidemics in Cuba in collaboration with the team of Dr Vivian Kourí Cardellá from IPK (Habana) [more details], both on the Cuban province level and on the world-wide one, to study the geographic origin of HIV introduction(s) to Cuba.
We are currently working on development of new methods for phylogeographic reconstruction and visualisation, allowing to perform reconstruction with a different level of detail in different parts of the tree, to incorporate spatial predictors [Lemey et al. PLOS Path 2014] in the maximum likelihood context, to formally compare phylogeographic predictions obtained with different tools and/or on slightly different phylogenies, and to reconstruct consensus scenarios.