Fig. 3

Clustering analysis of HHD-containing sequences. a The inclusion threshold (%identity) is incremented from top to bottom. Indicated percentages refer to inclusion thresholds where subclusters emerged, indicated by arrows. Outliers correspond to sequences that did not cluster successfully. The inclusion threshold is further incremented for subclusters with inhomogeneous sequence annotations. b Each resulting cluster is again submitted to an incrementation of the inclusion threshold. The first percentage value and arrow correspond to the limit identity where only one main cluster remains and point toward the number of sequences in that cluster. The second percentage value and arrows indicate the identity value where multiple clusters accounting for more than 10% of starting sequences emerge, pointing to the sizes of the two main clusters at the given identity threshold. c EFI—Enzyme Similarity Tool representation using a filter-value of 18. This threshold is determined empirically to differentiate and display all clusters in a single snapshot. The circled cluster highlights sequences from insects, corresponding to the Dyschronic sequences also circled in b