Fig. 2

The FRI workflow. 16S rRNA gene sequences are initially aligned against the reference sequences provided with Tax4Fun2 to identify the nearest neighbour. The OTU abundances for each sample are converted to presence-absence data and subsequently summarized based on the results from the nearest neighbour search. An association matrix (AM) containing the normalized functional profiles of those references identified in the 16S rRNA search is generated. The normalized functional profile in the AM is multiplied by the summarized OTU abundance. We assume that a function is present if the abundance of a function is greater than 50%. For instance, if 5 OTUs were affiliated to one 16S rRNA reference seqeunce, we would assume that every function with a minimum abundance of 10% in the normalized reference profile would be present in the AM. The FRI is finally calculated based on those community members possessing the function and their phylogenetic distribution. The amount of sequences/OTUs unused in the prediction (FSU/FTU) is provided in a log file 16S rRNA gene sequences are initially aligned against the reference sequences provided with Tax4Fun2 to identify the nearest neighbour. The presence of each OTU are summarized based on the results from the nearest neighbour search. An association matrix (AM) containing the functional profiles of those references identified in the 16S rRNA search is generated. The summarized abundances and the functional profiles stored in the AM are multiplied. Functions being present in all genomes are assumed to be present in the entire cluster. The functional redundancy is subsequently predicted from the average phylogenetic distance of all community members potentially possessing the function. This distance is subsequently normalized either by the average phylogenetic distance of the total reference tree (absolute) or by the distance found in the surveyed community (relative). FTU and FSU values are provided as a log file. The functions performing each step are marked with numbers