Example analysis of the primate MX1 gene. (A) FREEDA’s graphical user interface is divided into an input window (left half) and an output window (right half). The input window is used to provide a gene name (a), select the reference species (b), indicate where to save the data (c), and start the analysis (d). Optionally, the user can select advanced features (“Duplication expected,” “Tandem duplication expected,” “Long introns expected (>50 kb),” and “Common domains expected”; see online documentation: https://ddudka9.github.io/freeda/; e), label up to three regions of choice on the protein structure (f), select an additional codon frequency model (g), exclude species from the analysis (h), select a subgroup (i), and abort the analysis (j). The output window shows current tasks (“Events window,” top) and key results for each gene (“Results window,” bottom). The bottom part of the output window (green font) displays interactive messages guiding the user on how to provide input. (B) Putative adaptive sites are mapped onto the reference coding sequence. Graphs show recurrently changing residues (top, black bars), residues that are likely to have evolved under positive selection (middle, blue bars, probability ≥ 0.7), and most likely targets of positive selection (bottom, magenta bars, probability ≥ 0.9). Gray bars in all graphs show residues removed from the analysis. (C) Residues with the highest probability of positive selection (magenta) are mapped on the structural prediction model of the MxA protein (encoded by MX1) from AlphaFold. The N- and C-termini and known domains are automatically annotated, in addition to the user-specified region (“My motif of interest”). Regions removed from the analysis (arrowhead) are colored dark gray. To clearly show residues under positive selection, labels were modified manually in PyMOL (raw output is shown in Fig. S2 E).