Fig. 5: Topological features are enriched in damaging variants.
A The 3D structures of human ACE2 (top) and HBB (bottom), coloured by their per-residue two-dimensional TIFs. Structural analysis of missense variants for these genes predicts a number of them to be damaging44,45. Our topological analysis shows that amino acid substitutions causing structural damage are more likely to happen where the TIF is high, as shown by the violin plots on the right-hand site. This pattern is maintained across a dataset of disease-associated missense variants (B). As a further example, (C) shows human Adenylosuccinate lyase, with its missense variants highlighted on the 3D structure, and a plot of its 2-dimensional TIFs on the bottom. The n damaging and neutral variants for ACE2 are 49 and 200, for HBB are 81 and 396, and overall: 1418 and 6600, respectively. All box plots are shown with a box from the first to third quantile, median as a solid line, mean as a dashed line, and a line from the minimum to maximum value, excluding outliers. Source data are provided as a Source Data file in Source_data/source_data_Figure5A_ACE2.csv, Source_data/source_data_Figure5A_HBB.csv, and Source_data/source_data_Figure5B-S13-S14.csv.
