Microbial community diversity and network analysis
Rui Guan,Ruben Garrido Oter Max Planck Institute for Plant Breeding Research
Abstract
With the help of rapidly developing sequencing technologies, an increasing number of microbiome datasets have been generated and analysed. At present, analysis of taxonomic profiling data is mainly conducted using composition-based methods, which ignores interactions between community members and limits the study of community dynamics. To better understand the principles that govern the establishment of these microbial communities, we developed a framework for microbial community diversity analysis based on higher-order features. By integrating large quantities of available amplicon data, we inferred a large-scale co-occurrence network, from which diverse features such as clusters of microbes with co-varying abundances were extracted. Using variance analysis, we show that network-derived distance measures decrease the proportion of unexplained variance compared to traditional composition-based approaches, indicating their potential to extract novel insights from community profiling data. Furthermore, we introduced a bootstrap-permutation-based method to compare community networks at the global and local scales and to statistically assess their distances.
Keywords: microbiome,diversity,network