Supplementary Figure 2: Application of similarity regression to TFs with an array of DBDs.
From: Similarity regression predicts evolution of transcription factor sequence specificity
(a) DBDs are first aligned to find the best (maximizing amino acid identity) ungapped and internal alignment. Examples of permissible and non-permissable alignment configurations are shown. (b) Alignments are then scored by calculating positional protein similarity features in each finger of a DBD array (for example C2H2 ZFs), and combined into a single representation by averaging the features by the length of the longest DBD array.
