Microbiome: Data Normalization and Comparison of Bacterial Vaginosis Species Composition

Document Type

Presentation Abstract

Presentation Date

3-14-2016

Abstract

Human Microbiome Project (HMP) is a large scale nationwide study that utilizes next generation sequencing technology (NGS) to investigate the relationships between the human microbiota composition, diet and health status. We discuss common statistical challenges arising in microbiome data analysis and methods to approach these problems. One particular characteristic of these studies is that the data are often quite sparse but collected on a large number of variables, many of which are possible contaminants. First, we propose a network based data normalization method, which is known in microbiome literature as filtering. This method removes variables from a high dimensional sparse data set that are most likely present due to contamination. Then we discuss Co-Inertia Analysis approach to identify microbial species that contribute to major differences between healthy and bacterial vaginosis data. This disease, caused by excessive bacteria in vagina, affects approximately 30% of reproductive age women. The importance of this disease is that it has a high recurrence rate, and associated with miscarriage, preterm birth, and increased risk of acquiring other sexually transmitted infections.

Additional Details

Monday, March 14, 2016 at 3:10 p.m. in Math 103
Refreshments at 4:00 p.m. in Math Lounge 109

This document is currently not available here.

Share

COinS