Activity Number:
|
118
|
Type:
|
Topic Contributed
|
Date/Time:
|
Monday, August 1, 2016 : 8:30 AM to 10:20 AM
|
Sponsor:
|
Section on Statistics in Genomics and Genetics
|
Abstract #319443
|
View Presentation
|
Title:
|
Mediation Analysis of High-Dimensional Human Microbiome Data in the Longitudinal Study
|
Author(s):
|
Huilin Li * and Yilong Zhang and Martin J. Blaser
|
Companies:
|
New York University Langone Medical Center and New York University and New York University
|
Keywords:
|
High dimensional data analysis ;
Longitudinal study ;
Mediation analysis ;
Microbiome ;
Net-constraint ;
sparse PCA
|
Abstract:
|
Human microbiome can have a major impact on human health. Recent researches have been focused on causally linking the identified differences in the human microbiota with distinct human phenotypes including disease. However, the special data structure and characteristic of human microbiome data create challenges for the data analysis while combining the mediation analysis and dimension reduction method within the longitudinal study. In this paper, we provide a framework to test and estimate the microbiome mediation effect and identify key taxa efficiently by using all of the longitudinal data. We first propose a phylogenetic tree constrained sparse principle analysis method to reconstruct the microbiome profile with the selected key taxa. We then integrate the selected taxa into the linear mixed model to test and estimate the microbiome mediation effect. We evaluated our method through extensive simulations and illustrate our method through a longitudinal murine microbiome study.
|
Authors who are presenting talks have a * after their name.