Subsequently, the bioinformatic procedures for each paper were systematically parsed to identify the different tasks implemented, i.e. specific bioinformatic actions with a clearly-defined purpose and performed by a single tool. A total of 30 distinct bioinformatic tasks were identified starting from initial procedures on raw sequencing files through to the generation of community tables (see Table 1 for a description of each task). We focused solely on bioinformatic tasks that were presented as necessary for the generation of information about the occurrence or incidence of taxonomic units in the sampled communities (i.e. community data), and the taxonomic identity of these units. For example, we did not record any steps performing phylogenetics with a final OTU set, although we recorded steps where phylogeny-based methods were used as part of OTU delimitation and filtering. Similarly, we recorded tasks that performed filtering of community data for the purposes of removing OTUs or OTU records arising from erroneous sequences or from cross-talk/contamination (Edgar, 2018), but we did not record tasks that filtered community data for the purposes of statistical correction, such as normalisation or rarefaction.