Fig. 7

a-c The importance of genes is visualized in three plots based on a random forest analysis. For continuous variables (age at onset, C9orf72 expansion size, and poly[GP] levels), the importance is defined as an increase in mean squared error. The blue gradient represents the importance of each gene, from very important (light) to less important (dark). Arrows point at genes of interest, namely VEGFA, CDKL1, and EEF2K (Table 7 and Fig. 6)