Figure 4.

Heatmap of KO-Numbers collated into relevant functions associated with detected orthologues. OGs were predicted from the co-assembled metagenomes with the eggNOG Mapper (Huerta-Cepas et al., 2019) and their abundance was calculated by mapping the information of the corresponding encoding contigs back to the reads. Displayed is the relative abundance of the orthologues in each sample, normalized to the abundance of the universal housekeeping gene/function COG0468 (recA). In cases where closely related OGs were grouped into more general metabolic functions, the corresponding mean value is shown. Functional groups with a total relative abundance below 0.02 have been removed for better representability.