Approximating Genealogies for Partially Linked Neutral Loci Under a Selective Sweep

Consider a genetic locus carrying a strongly beneficial allele which has recently fixed in a large population. As strongly beneficial alleles fix quickly, sequence diversity at partially linked neutral loci is reduced. This phenomenon is known as a selective sweep. The fixation of the beneficial allele not only affects sequence diversity at single neutral loci but also the joint allele distribution of several partially linked neutral loci. This distribution can be studied using the ancestral recombination graph for samples of partially linked neutral loci during the selective sweep. To approximate this graph, we extend recent work by Etheridge et al. (Ann Appl Probab 16:685-729, 2006) and Schweinsberg and Durrett (Ann Appl Probab 15:1591-1651, 2005) using a marked Yule tree for the genealogy at a single neutral locus linked to a strongly beneficial one. We focus on joint genealogies at two partially linked neutral loci in the case of large selection coefficients alpha and recombination rates rho = theta(alpha/log alpha) between loci. Our approach leads to a full description of the genealogy with accuracy of theta((log alpha)(-2)) in probability. As an application, we derive the expectation of Lewontin's D as a measure for non-random association of alleles.