Simultaneous Statistical Inference for Epigenetic Data

摘要: Epigenetic research leads to complex data structures. Since parametric model assumptions for the distribution of epigenetic are hard verify we introduce in present work a nonparametric statistical framework two-group comparisons. Furthermore, analyses often performed at various genetic loci simultaneously. Hence, order be able draw valid conclusions specific loci, an appropriate multiple testing correction is necessary. Finally, with technologies available simultaneous assessment many interrelated biological parameters (such as gene arrays), approaches also need deal possibly unknown dependency structure data. Our approach comparison two samples independent multivariate observables based on recently developed permutation tests. We adapt their theory cope families hypotheses regarding relative effects. results indicate that test keeps pre-assigned type I error level global null hypothesis. In combination closure principle, family-wise rate corresponding locus/parameter-specific can controlled. applications demonstrate group differences detected reliably our methodology.