Title: A new R library for discriminate groups based on abundance profile and biodiversity on microbiome metagenomics matrices
Authors: Clara Rodriguez-Casado - Section of Statistics Fac Biology University of Barcelona (Spain)
Jorge Frias-Lopez - Department of Microbiology - Forsyth Institute (United States)
Toni Monleon-Getino - University of Barcelona (Spain) [presenting]
Abstract: Metagenomics is the study of genetic material recovered directly from samples like microbiome which have shown an extraordinary diversity. While traditional microbiology and microbial genome sequencing rely upon cultivated clonal cultures, metagenomics gene sequencing cloned specific genes to produce a profile of diversity in the microbiome. There is a great interest in associating specific groups of organisms with health and disease typologies, but unfortunately there are very few statistical tools for profile analysis with the ability to differentiate types which can co-exist. We have developed MetagenOutlineLDA, a new library for R that allows a statistical analysis of metagenomic matrices using different statistical approaches. This library performs three basic tasks: 1) the estimation of metagenomic abundance profiles (relative abundance of species) for each sample using robust regression, 2) estimation of metagenomic biodiversity-alpha and 3) performing a discriminant analysis (LDA,SVM,NN, etc.) to distinguish groups (e.g. healthy /disease). Thus a new metagenomic analysis could indicate the group of belonging with a reasonable percentage. Examples of analysis using MetagenOutlineLDA with people affected by periodontitis and Crohn's disease are presented proving to be a useful library. The results confirm that the diseases studied not only alter the composition of the human microbiome, but also its structure.