|
Activity Number:
|
244
- Statistical methods for microbiome data analysis and beyond
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 11, 2021 : 10:00 AM to 11:50 AM
|
|
Sponsor:
|
Section on Statistical Consulting
|
|
Abstract #317682
|
|
|
Title:
|
Statistical Analysis of Longitudinal Microbiome Data
|
|
Author(s):
|
Amy Pan* and Vy Lam and Samantha N. Atkinson and Nita Salzman and L Silvia Munoz-Price
|
|
Companies:
|
Medical College of Wisconsin and PITA Analytics and Medical College of Wisconsin and Medical College of Wisconsin and Medical College of Wisconsin
|
|
Keywords:
|
Microbiome;
Longitudinal;
Sparsity;
Overdispersion;
Zero-inflation;
Analysis
|
|
Abstract:
|
Introduction. The unique feature and complex microbiome data from high-throughput DNA sequencing, especially the sparsity and overdispersion of the data, present challenges to statistical analysis and interpretation. Longitudinal data adds more complexity to the existing challenges. Methods. Fecal samples were collected over time in patients who have tested at least twice for the presence of Clostridium difficile. Patients were classified as the following groups: negative-to-positive; negative controls. 16S rRNA gene sequencing was used to examine taxonomic composition and species diversity of the bacteria community. Results. Baseline values and slopes were compared between groups (negative-to-positive vs negative controls) for genera that might be associated with the presence of C. difficile. A zero-inflated negative binomial mixed effects model fit the data better than a zero-inflated Poisson mixed effects model or a generalized mixed effects model. Conclusions. Zero-inflated negative binomial mixed effects model tackles sparsity and overdispersion while addressing time trends and within-subject correlations.
|
Authors who are presenting talks have a * after their name.
