Effects of Zeros on Analysis in the Microbiome Data

• Activity Number: 535 - Contributed Poster Presentations: Section on Statistics in Genomics and Genetics
• Type: Contributed
• Date/Time: Wednesday, August 1, 2018 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistics in Genomics and Genetics
• Author(s): Amy Pan* and Nita Salzman and T.Hang Nghiem-Rao and Martin Hessner and Pippa Simpson
• Companies: Medical College of Wisconsin and Medical College of Wisconsin and Medical College of Wisconsin and Medical College of Wisconsin and Medical College of Wisconsin
• Keywords: microbiome; OTUs; zeros

Abstract

Introduction. Microbiome exists in large amounts in human beings. It is likely that human microbiomes are involved in the pathogenesis of many diseases. Despite intense investigation in human microbiome research, the statistical methods for microbiome data are still being developed. 16S ribosomal RNA (rRNA) gene targeted amplicon sequencing is commonly used to examine taxonomic composition and species diversity of the bacteria community. Operational Taxonomic Units (OTUs) are obtained via read error correction, filtering, clustering at 97% sequence similarity. Methods. OTUs are often skewed and over dispersed and may be better represented by a mixture of distributions instead of one. Furthermore, many OTUs are dominated by zeros. To investigate the effect of zeros on analyses in the microbiome data, simulation studies are used to find the optimal statistical method to compare two treatment groups when different proportion of zeros present in the data. Results. The optimal statistical methods are different for different proportions of zeros present in the data.