2.6. PICRUSt Analysis

This analysis relies on the obtained table of operational taxonomic units (OTUs) for each sample against the Greengenes ribosomal RNA database (release of August 2013) at 97% similarity [41,42]. The obtained OTUs were normalized by dividing their abundances by known or predicted 16S rDNA gene copy number abundances before their final predictions, and the resultant KOs was used to express the functional information of the microbial community. The predicted functional counts were rarefied to the same depth for further analysis. KEGG orthologous gene data refers to the equivalent genes and gene products (e.g., RNA and proteins) that were produced in different organisms. Moreover, they contained subgroups of cellular processes, environmental information processing, genetic information processing, metabolism, organismal systems, and human diseases (Level 1). The accuracy of function predictions depends on how closely the obtained microbes in a given sample are related to the microbes with sequenced genome representatives, as measured by the Nearest Sequenced Taxon Index (NSTI) where the lower values indicate a closer mean relationship [26]. In addition, KEGG Orthology terms of the microbial communities predicated by PICRUSt were further analyzed by STAMP [40].