The mean and variance of the number of segregating sites since the last hitchhiking event

 Tight linkage may cause a reduction of nucleotide diversity in a chromosomal region if an advantageous mutation appears in that region which is driven to fixation by directional selection. This process is usually called genetic hitchhiking. If selection is strong, the entire process takes place during a time period of length 2s ln (2N) that is very short relative to 2N generations s is the selection coefficient of the advantageous mutation and N the effective diploid population size]. On the time scale of 2N generations, which is characteristic for neutral evolution, we may therefore call this process a hitchhiking event. Using coalescent methods, we analyzed a model in which a hitchhiking event occurred in a chromosomal region of zero-recombination in the past at time x. Such a hitchhiking “catastrophe” wipes out completely genetic variation that existed in a population before that time. Standing variation observed at present must therefore be due to mutations that have arisen since time point x. Assuming that all newly arising mutations are neutral, we derived expressions for the expectation, variance and also for the higher moments of the number of nucleotide sites segregating in a sample of n genes as a function of x. The result for the first moment is then used to estimate the time back to the last hitchhiking event based on DNA polymorphism data from Drosophila. Assuming that directional selection is the sole determinant of the level of genetic variation in the gene regions surveyed, we obtained estimates of x that were typically in the order of 0.1N generations. Received 14 May 1996; received in revised form 26 August 1996