Online Program

Investigation of False Discovery Rate for Microbiome Data from Unreplicated Block Design (309971)

Keywords: false discovery rate, microbiome data, unreplicated block design, compositional data

Microbiome data that characterize human (or animal) health and disease have been gaining enormous popularity among scientists. While novel statistical research is active in this field, easy access is a method from such as differential expression in genomics. The fundamental difference, however, is that the differential abundance in microbiome means the difference in relative abundances based on taxon counts, needing special attention to the “compositionality.” With such data, controlling false discovery rate has proven difficult, we aim to examine the well-known Benjamini and Hochberg (1995) procedure under unreplicated block design. For compositional microbiome data, it is useful to use the nonparametric test to avoid the distributional assumption. For paired data, Wilcoxon-signed rank test, Quade test, Sign test and Friedman test can be alternatives to the paired t-test. After some simulations regarding the assumptions from BH procedure, we found the Wilcoxon-signed rank and Quade test are appropriate for the FDR-based inference. Along with our suggestion for this specific methods, we demonstrate the key assumptions when using FDR procedure with microbiome data.