Fig. 7

Analysis of OrthoFiller on different subsets of plant species. Species subsets are indicated by bullets on the bottom panel. Behaviour of OrthoFiller is consistent across different numbers and relatedness of species, with the exception of when only two species are used. a, e Number of genes removed from A. thaliana (light + darker colour), and the proportion recovered (darker colour). b, f Proportion of genes found which overlap removed genes (darker colour) is consistently higher for OrthoFiller than for its de novo counterpart. c, g Percentage of recovered genes of low quality, i.e. with protein F-score <0.95, is consistently lower in OrthoFiller than the de novo version. d, h Mean protein F-score of lower-quality genes is consistently higher for OrthoFiller than for the de novo version