A New Proof of the Expected Frequency Spectrum under the Standard Neutral Model

The sample frequency spectrum is an informative and frequently employed approach for summarizing DNA variation data. Under the standard neutral model the expectation of the sample frequency spectrum has been derived by at least two distinct approaches. One relies on using results from diffusion approximations to the Wright-Fisher Model. The other is based on Pólya urn models that correspond to the standard coalescent model. A new proof of the expected frequency spectrum is presented here. It is a proof by induction and does not require diffusion results and does not require the somewhat complex sums and combinatorics of the derivations based on urn models.