Genealogies in Expanding Populations

TL;DR

This paper rigorously analyzes genealogies in a one-dimensional biased voter model, showing convergence to Wright-Fisher SPDEs and describing tracer dynamics with a new duality approach.

Contribution

It proves convergence of the biased voter model to Wright-Fisher SPDEs and introduces a novel duality equation for tracer dynamics.

Findings

Convergence of the biased voter model to Wright-Fisher SPDEs.

Description of genealogies as a branching Brownian motion without simultaneous coalescences.

Development of a new duality equation for tracer dynamics.

Abstract

The goal of this paper is to prove rigorous results for the behavior of genealogies in a one-dimensional long range biased voter model introduced by Hallatschek and Nelson [25]. The first step, which is easily accomplished using results of Mueller and Tribe [38], is to show that when space and time are rescaled correctly, our biased voter model converges to a Wright-Fisher SPDE. A simple extension of a result of Durrett and Restrepo [18] then shows that the dual branching coalescing random walk converges to a branching Brownian motion in which particles coalesce after an exponentially distributed amount of intersection local time. Brunet et al. [8] have conjectured that genealogies in models of this type are described by the Bolthausen-Sznitman coalescent, see [39]. However, in the model we study there are no simultaneous coalescences. Our third and most significant result concerns "tracer dynamics" in which some of the initial particles in the biased voter model are labeled. We show that the joint distribution of the labeled and unlabeled particles converges to the solution of a system of stochastic partial differential equations. A new duality equation that generalizes the one Shiga [44] developed for the Wright-Fisher SPDE is the key to the proof of that result.