Fig. 6

Trends in gene structure show hemipteroid-specific tendencies. a Median values per species for protein size, exon size, and exon number for a curated set of highly conserved genes encoding large proteins of diverse functional classes (see also Additional file 1: Supplemental Note 6.3). Sample sizes are indicated, with 11 genes for which orthologs were evaluated in all species. Where it was not possible to analyze all 30 genes for a given species, equal sampling was done across the range of protein sizes of the complete dataset, based on the Cimex ortholog sizes (1:1:1 sampling from big-to-medium-to-small subcategories of 10 genes each). b Box plot representations of coding sequence exon size (aa) for 2 species from each of 3 insect orders, based on datasets of unique coding sequence exons (1 isoform per gene) and excluding terminal exons < 10 aa (as most of those exons may rather be UTRs or a small placeholder N-terminal exon based on automated Maker model predictions). Only manually curated gene models were considered for the i5K species, including Oncopeltus; the entire OGS was used for Tribolium and Drosophila. For clarity, outliers are omitted; whiskers represent 1.5× the value of the Q3 (upper) or Q2 (lower) quartile range. MAD, median absolute deviation. Species are represented by their 4-letter abbreviations, with their ordinal relationships given below the phylogeny in a: Hemip., Hemiptera; Thys., Thysanoptera; Col., Coleoptera; Dipt., Diptera. Species abbreviations as in Figs. 2 and 4 and additionally Gbue, Gerris buenoi [164]; Agla, Anoplophora glabripennis [30]; Ccap, Ceratitis capitata [165]