Extended Data Fig. 1: Illustrating the correlation between BGC clustering thresholds and the grouping of their pathway products.
From: Compendium of specialized metabolite biosynthetic diversity encoded in bacterial genomes
a) a snippet of a complete-linkage hierarchical dendrogram constructed by doing a pairwise distance comparison of L2-normalized BGC features within the MIBiG dataset, highlighting the grouping of BGCs for the enediynes Uncialamycin (UCM) and Tiancimycin (TNM) under the threshold T = 0.5, and further grouping with another related enediyne BGC, Dynemicin (DNM) under the looser threshold of T = 0.7. b) Comparative genes analysis generated using the clinker tool92 v0.0.23 shows how UCM and TNM BGCs are much more similar to each other than to DNM (same-colored genes indicate <70% amino acid similarity, while colored edges indicate <50% amino acid similarity), which is consistent with the structural diversity of their compounds (pictured).
