Activity Number:
|
35
- Epidemiological Models for Genetic Data, Biomarkers, and Rare Outcomes
|
Type:
|
Contributed
|
Date/Time:
|
Sunday, August 7, 2022 : 2:00 PM to 3:50 PM
|
Sponsor:
|
Section on Statistics in Epidemiology
|
Abstract #322236
|
|
Title:
|
A Maximum-Type Multivariate Microbial Differential Abundance Test with Application to Microbiome Studies
|
Author(s):
|
Jiyuan Hu* and TingFang Lee and Zhengbang Li
|
Companies:
|
NYU Grossman School of Medici and NYU Grossman School of Medicine and Central China Normal University
|
Keywords:
|
High dimensional;
microbiome data;
compositional;
differential abundance test
|
Abstract:
|
Compositional data is commonly encountered in many statistical application fields. How to efficiently test the difference between two high-dimensional compositional data remains theoretically and computationally challenging. In this study, we propose a novel test for identifying the mean difference of two high-dimensional compositional mean vector based on the centered log-ratio transformation of the compositions. Our proposed test is free of the equal covariance matrix assumption, therefore provides more flexibility. The asymptotic null distribution is obtained and the asymptotical power against sparse alternatives is also investigated. Simulation results show that the proposed test can control empirical sizes well no matter two covariance matrices are equal or not. The usefulness of the proposed tests is further illustrated by two real data analyses.
|
Authors who are presenting talks have a * after their name.