Binary Galton-Watson trees with mutations

Abstract: We consider a family of marked binary Galton-Watson trees that can model individual types in population genetics, by allowing for mutation and reversion in discrete and continuous time. We derive a recursive formula for the computation of the joint distribution of types conditional to the value of the total progeny. This allows us to compute the evolution of various expected quantities, such as the mean proportions of different types as the tree size or time increases. Some generating functions are determined in explicit forms using generalized Catalan numbers, and integrability criteria are obtained as a consequence.