A Novel Normalization and Differential Abundance Test Framework for Microbiome Data

• Activity Number: 40 - Statistical Methods for Microbiome and Tumor Data
• Type: Contributed
• Date/Time: Sunday, July 28, 2019 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistics in Genomics and Genetics
• Author(s): Yuanjing Ma* and Yuan Luo and Hongmei Jiang
• Companies: and Northwestern University and Northwestern University

Abstract

We develop a novel framework for differential abundance analysis on sparse high-dimensional marker gene microbiome data. Our methodology relies on a network-based normalization technique and a two stage zero-inflated mixture count regression model (RioNorm2). Our novel network-based normalization method aims to find a group of relatively invariant species across samples and environments in order to construct size factors. It does not make any assumption on count distributions. Another contribution of the paper is that our testing approach can take into consideration under-sampling and over-dispersion with flexibility by separating microbiome species into different subgroups and model them separately. Through comprehensive simulation studies, the performance of our method is consistently powerful and robust across different settings with different sample sizes, library sizes and effect sizes. We also demonstrate the effectiveness of our novel framework using a published dataset of Metastatic Melanoma and find biological insights from the results.