Microbiome Normalization Methods: Effect on Ordination Analysis

• Activity Number: 119
• Type: Topic Contributed
• Date/Time: Monday, August 1, 2016 : 8:30 AM to 10:20 AM
• Sponsor: International Indian Statistical Association
• Author(s): Ekaterina Smirnova* and Snehalata Huzurbazar and Glen Alan Satten and Liyang Diao
• Companies: University of Wyoming and University of Wyoming and CDC and Yale University
• Keywords: microbiome ; small-n large-p ; data normalization ; RNA-seq ; ordination

Abstract

Microbiome next generation sequencing experiments measure counts of DNA fragments for a large number of species in a sample. The total number of reads might vary dramatically across samples, and therefore appropriate scaling is necessary for any analysis. Prior to data analysis, normalization and bias adjustment are often implemented using filtering, library size adjustment and variance stabilization methods. Methods developed for RNA-seq data are being used for microbiome data, however the latter are extremely sparse and often dominated by a small number of species. Using methods from RNA-seq literature, and recently developed methods for adjustment of microbiome data, we investigate the effects of normalization and bias adjustment on ordination based analyses.