Nonlinear machine learning pattern recognition and bacteria-metabolite multilayer network analysis of perturbed gastric microbiome

The stomach is inhabited by diverse microbial communities, co-existing in a dynamic balance.Long-term use of drugs such as proton pump inhibitors (PPIs), or bacterial infection such asHelicobacter pylori, cause significant microbial alterations. Yet, studies revealing how thecommensal bacteria re-organize, due to these perturbations of the gastric environment, are inearly phase and rely principally on linear techniques for multivariate analysis. Here we dis-close the importance of complementing linear dimensionality reduction techniques withnonlinear ones to unveil hidden patterns that remain unseen by linear embedding. Then, weprove the advantages to complete multivariate pattern analysis with differential networkanalysis, to reveal mechanisms of bacterial network re-organizations which emerge fromperturbations induced by a medical treatment (PPIs) or an infectious state (H. pylori). Finally,we show how to build bacteria-metabolite multilayer networks that can deepen our under-standing of the metabolite pathways significantly associated to the perturbed microbialcommunities.