Pedigree in the biparental Moran model (Q2140017)

This paper is concerned with the genetic composition of a population with biparental genetic transmission. The authors focus on the Moran biparental model, in which 1) each individual has 2 parents, who contribute equally to the genome of their offspring, 2) both parents and the replaced individual are in different sites and 3) the number of individuals is fixed, given by \(N\). The weight of an ancestor is defined to be the probability that a given gene of the sampled individual comes from this ancestor. Under this idealized model, the authors first show that the weights of an ancestor in all individuals are asymptotically equal. As \(N\) goes to infinity, the total weights of \(l\) ancestors in the population (properly rescaled by a factor of \(N\)) converge in law to a vector of \(l\) independent random variables, that are the mixture of a Dirac measure in 0 and an exponential law with parameter 1/2. The key points of the proofs are 1) introducing the pedigree graph of the population, 2) Considering the stationary distribution of a \(k\)-particle random walk on the random pedigree graph and 3) a certain projection of this \(k\)-particles random walk remains a Markov chain.