Detecting differentially methylated regions using a fast wavelet-based approach to functional association analysis

摘要

Comparison of the computational time under H0 when performing an EWAS with FFW and WaveQTL for different sample sizes. The y-axis has been log-transformed. The green curve indicates the computational time using the permutation procedure described by Besag and Clifford [29], which is based on a maximum of 10 permutations being larger than the observed test statistics (the set up is according to Shim and Stephens [8]). The orange curve indicates the computational time using FFW. The blue line indicates the computational time of the simulations required in FFW, which remains constant with increasing sample size