Comparison of and . An example for the comparison of performances between and , where f = 1%, αstart = 1%, n = 3000, K = 100, and G = 10. The two box plots on the left are a comparison between and as an estimate of f and , where shows a considerable tendency of over-estimation compared to . The upper right histogram shows how deviates from f in 1000 simulations, which appears to be roughly unbiasedly distributed around f. The lower right bar plot is a summary plot of the MSE for these two methods as estimates of f and ffrac, showing superiority of both as an estimator of f and of ffrac, in terms of MSE and Cg.
Chen and Sun BMC Genomics 2013 14(Suppl 1):S1 doi:10.1186/1471-2164-14-S1-S1