Fig. 3

Enzyme annotation performance comparison and example outputs. A Stacked barcharts indicate the number of enzymes, as defined through enzyme classification (EC) assignments, annotated by each pipeline for the three sets of datasets: NOD mouse gut, kimchi, and human oral biofilm. In addition to displaying ECs’ unique or shared between MetaPro and the other two tools, also shown are ECs, predicted by HUMAnN3 and SAMSA2 to occur in combination with another EC, in the same transcript, with no supporting evidence that such a combination has been previously observed (as defined through Swiss-Prot annotations). Further, for HUMAnN3, we show the number of EC assignments that occur in combinations of greater than 2 ECs. B A Cytoscape network representation of the tricarboxylic acid (TCA) cycle, together with the breakdown of EC expression by taxon. MetaPro generates a Cytoscape compatible annotation file which can be used to map gene expression data onto KEGG defined pathways using the KEGGscape and enhancedGraphics plugin applications. Here, each node represents an individual enzyme with its size indicating the overall expression of that enzyme in the dataset (as defined by RPKM). Colored sectors indicate the contribution of each taxon to expression of that enzyme. To simplify the display, taxa were manually merged into 8 taxonomic groupings. Here, we see that members of Klebsiella and Citrobacter are the main contributors to the TCA cycle. These outputs were generated from a sample in the human oral biofilm dataset (SRR5984039). C A summary overview showing the contribution of major taxa to KEGG-defined superpathways. After annotation of enzymes, MetaPro generates a heatmap in PNG format showing taxa responsible for expression (calculated as the sum of RPKMs assigned to each EC annotated for each taxon). Only taxa associated with at least 1% of total reads are shown (see the “Methods” section)