Dating Rare Mutations from Small Samples with Dense Marker Data

We present a method for estimating the age of a mutation based on the genetic length of ancestral haplotypes shared between individuals carrying the mutation. The method can be reliably applied to small samples, typical of situations involving rare mutations, and makes effective use of modern high-density SNP data, thus overcoming two of the limitations with existing methods. The method provides age estimates and confidence intervals without the use of asymptotic theory and is applicable to genealogies in which the data are independent or correlated. In the correlated case we estimate the correlation directly from the data, rather than relying on a model for the genealogy. To demonstrate the method's efficacy, we provide simulation results and compare it to other methods. The length data are obtained with a simple procedure, and an R script is available for performing the calculations.