Estimation of Selection Parameters using Multi-generation Cytonuclear Data

Testing Neutrality of mtDNA Using Multigeneration Cytonuclear Data

“…Therefore, a test of neutrality can be based on the score statistic s(wo), with wo = (1/4, , 1/4). Since s, the derivative of the log-likelihood based on the population gametic relative frequencies, is not computable from the observed data, Datta (2001) suggested using an approximate version of it obtained by replacing them with the corresponding sample versions. This results in additional terms for the variance-covariance matrix, all of which can be consistently estimated from the observed data.…”

“…See Datta (2001) for the details of the algebraic calculations. Of course, the approximate score is defined as s(w) ^ (d/dw) I (w) where we interpret it as a function of w = (wι,W2, ws) ( i.e., replace w^ by I -wι -w 2 -w^).…”

“…Of course, the approximate score is defined as s(w) ^ (d/dw) I (w) where we interpret it as a function of w = (wι,W2, ws) ( i.e., replace w^ by I -wι -w 2 -w^). Datta (2001) was able to calculate the estimated asymptotic variancecovariance matrix of s given by Σy = (d/dw)(s(w))\ w=g…”

“…We now report the results of a simulation study where we compare the power of Datta's (2001) test with those of earlier tests by Schaffer et al (1977). The experimental setup is as follows.…”

“…As we will see in Section 3, the existing statistical tests are often too limited to take full advantage of the multi-generation cytonuclear data that are now available. As a result, a new test based on the recent works by Datta (1999Datta ( , 2001 is proposed. This test is based on an approximate likelihood for the full available data constructed from a broad parametric selection model and is therefore expected to perform well in practice.…”