Activity Number:
|
394
- Brushing up Your Skills in Genomic Data Analysis
|
Type:
|
Topic Contributed
|
Date/Time:
|
Tuesday, July 30, 2019 : 2:00 PM to 3:50 PM
|
Sponsor:
|
Korean International Statistical Society
|
Abstract #307115
|
|
Title:
|
Statistical Considerations for Metabolomic Data
|
Author(s):
|
Sharon Lutz* and Rachel S. Kelly and Joanne E. Sordillo and Ann Wu
|
Companies:
|
Harvard Medical School and Channing Division of Network Medicine, Brigham & Women’s Hospital, Harvard Medical School and Harvard Medical School and Harvard Pilgrim Health Care and Harvard Medical School and Harvard Pilgrim Health Care
|
Keywords:
|
Metabolomics;
longitudinal data;
linear mixed models
|
Abstract:
|
While germline DNA does not vary as a function of time, the metabolome can vary substantially over a given time. Disease progression and severity for such common diseases as asthma can also change substantially over time. Given this time dependent nature, we consider a generalization of standard mixed models for longitudinal data that include flexible mean functions as well as combined autoregressive (AR) and compound symmetry (CS) covariance structures. Through simulation studies, we compare this approach to standard regression approaches and consider issues associated with metabolomic data in this context. We apply these methods to the Childhood Asthma Management Program (CAMP) study and the Genetic Epidemiology of Asthma in Costa Rica Study (GACRS) to determine the role of the metabolome in modifying the association between age and bronchodilator response (BDR) in individuals with asthma.
|
Authors who are presenting talks have a * after their name.