Searcharxiv⌕ Search

arXiv subjects

Éric Brunet

Publications and source records attributed to Éric Brunet.

23 records · Page 2Linked to original sources

Growth rates of the population in a branching Brownian motion with an inhomogeneous breeding potential

We consider a branching particle system where each particle moves as an independent Brownian motion and breeds at a rate proportional to its distance from the origin raised to the power $p$, for $p\in[0,2)$. The asymptotic behaviour of the right-most particle for this system is already known; in this article we give large deviations probabilities for particles following "difficult" paths, growth rates along "easy" paths, the total population growth rate, and we derive the optimal paths which particles must follow to achieve this growth rate.

math.PR↗

Accessibility percolation with backsteps

Consider a graph in which each site is endowed with a value called \emph{fitness}. A path in the graph is said to be "open" or "accessible" if the fitness values along that path is strictly increasing. We say that there is accessibility percolation between two sites when such a path between them exists. Motivated by the so called House-of-Cards model from evolutionary biology, we consider this question on the $L$-hypercube $\{0,1\}^L$ where the fitness values are independent random variables. We show that, in the large $L$ limit, the probability that an accessible path exists from an arbitrary starting point to the (random) fittest site is no more than $x^*_{1/2}= 1-\frac12\sinh^{-1}(2) =0.27818\ldots$ and we conjecture that this probability does converge to $x^*_{1/2}$. More precisely, there is a phase transition on the value of the fitness $x$ of the starting site: assuming that the fitnesses are uniform in $[0,1]$, we show that, in the large $L$ limit, there is almost surely no path to the fittest site if $x>x^*_{1/2}$ and we conjecture that there are almost surely many paths if $x<x^*_{1/2}$. If one conditions on the fittest site to be on the opposite corner of the starting site rather than being randomly chosen, the picture remains the same but with the critical point being now $x^*_1= 1-\sinh^{-1}(1)= 0.11863\ldots$. Along the way, we obtain a large $L$ estimation for the number of self-avoiding paths joining two opposite corners of the $L$-hypercube.

math.PR↗

Genealogies in simple models of evolution

We review the statistical properties of the genealogies of a few models of evolution. In the asexual case, selection leads to coalescence times which grow logarithmically with the size of the population in contrast with the linear growth of the neutral case. Moreover for a whole class of models, the statistics of the genealogies are those of the Bolthausen-Sznitman coalescent rather than the Kingman coalescent in the neutral case. For sexual reproduction, the time to reach the first common ancestors to the whole population and the time for all individuals to have all their ancestors in common are also logarithmic in the neutral case, as predicted by Chang []. We discuss how these times are modified in a simple way of introducing selection.

q-bio.PE↗

How genealogies are affected by the speed of evolution

In a series of recent works it has been shown that a class of simple models of evolving populations under selection leads to genealogical trees whose statistics are given by the Bolthausen-Sznitman coalescent rather than by the well known Kingman coalescent in the case of neutral evolution. Here we show that when conditioning the genealogies on the speed of evolution, one finds a one parameter family of tree statistics which interpolates between the Bolthausen-Sznitman and Kingman's coalescents. This interpolation can be calculated explicitly for one specific version of the model, the exponential model. Numerical simulations of another version of the model and a phenomenological theory indicate that this one-parameter family of tree statistics could be universal. We compare this tree structure with those appearing in other contexts, in particular in the mean field theory of spin glasses.

cond-mat.dis-nn↗

A branching random walk seen from the tip

We show that all the time-dependent statistical properties of the rightmost points of a branching Brownian motion can be extracted from the traveling wave solutions of the Fisher-KPP equation. We show that the distribution of all the distances between the rightmost points has a long time limit which can be understood as the delay of the Fisher-KPP traveling waves when the initial condition is modified. The limiting measure exhibits the surprising property of superposability: the statistical properties of the distances between the rightmost points of the union of two realizations of the branching Brownian motion shifted by arbitrary amounts are the same as those of a single realization. We discuss the extension of our results to more general branching random walks.

cond-mat.stat-mech↗