Co-occurrence analysis of bacterial families using metagenomic data and network analysis

ORLOWSKI J.F.; CORREA O.; MONTECCHIA M.S.; SORIA M.A.

Rosario, Santa Fe

Congreso; IV Congreso Argentino de Bioinformática y Biología Computacional (4CAB2C) & IV Conferencia de la Sociedad Iberoamericana de Bioinformática (SOIBIO); 2013

Asociación Argentina de Bioinformática y Biología Computacional (A2B2C) & Sociedad Iberoamericana de Bioinformática (SOIBIO)

The massive sequencing of amplicons using next-generation technologies has had a deep impact in soil microbial ecology. A significant number of new taxa were discovered and many different environments and conditions were surveyed. Part of the usual analytical pipeline is the search of core microbiomes and rare taxa. While the search of core microbiomes is useful to determine the composition of the shared communities, they are aggregated data that miss the fine structure of the patterns of taxa co-occurrence. The application of network analysis has been proposed as an alternative way to study patterns of co-occurrence at a higher level of detail. In this work we extend the application of network analysis to a meta-analysis of several datasets and determine the structure of a highly connected microbiome applying network clustering techniques. We characterized a meta-community derived from five different datasets of 16S rRNA amplicon data applying network analysis methods at the family taxonomical level. The data matrix contained 210 bacterial families and 190 samples. The analysis of the derived networks showed the presence of local bacterial families and two groups of connected families. The first group was composed of low frequency families present in several, but not most, samples. The second group included high frequency families with simultaneous presences in many environments. It includes members from the Corynebacteriaceae, Bacillaceae, Burkholderiaceae, Desulfovibrionaceae, Sphingobacteriaceae, Xanthomonadaceae and Sphingomonadaceae families, and from the Actinomycetales and Sphingobacteriales orders. This group of highly connected families could be further separated into three communities. The flexibility and variety of tools for the analysis of networks open the door to a more thorough analysis of community structure.