FRESCo demonstrates high specificity in tests on simulated regions of excess synonymous constraint. (A) On a simulated dataset of 1,000 sequences with regions of varying strength of synonymous constraint, FRESCo recovers SCEs with high accuracy. We plot the synonymous substitution rate at a 10-codon resolution, displaying below the plot the relative synonymous substitution rate in each portion of the sequence. The red tracks at the bottom show recovered regions of significant excess synonymous constraint at window sizes of 1, 5, 10, 20, and 50 codons. (B) Recovery of simulated regions of excess synonymous constraint improves with increasing branch length (in substitutions/site), strength of synonymous constraint, and number of aligned sequences (5-codon sliding windows). (C) Distribution of P-values in simulated sequence where there is no synonymous constraint. Q-Q plots of the distribution of P-values for 5-codon sliding windows in simulations based on alignments of 100 (top), 500 (middle), and 1,000 (bottom) random sequences. Each plot is based on 20 independent, 500-codon simulated alignments (total of 10,000 codons).