Figure 1. Likelihood-ratio test on short linear motifs after gene duplication on simulated data.

A) Schematic of the motif-specific likelihoodratio test applied to all motifs. Rates of evolution are computed for each motif before (α_pre-WGD) and after (α_WGD) gene duplication and compared with the rates that were observed for the whole protein (see Methods). Red double arrow illustrates the duplication event. Bolded clades are clades with significant changes in constraints. Striped patterned boxes indicate short linear motifs with significantly different rate of evolution. D_KL indicates the expected deviation of the likelihood-ratio test from the whole protein. B) Alignment of the N-terminus of the Dbp1/Ded1 homologs illustrates the rate heterogeneity amongst columns and highlights the short length of a putative motif (black rectangle zoom). Blue shade represents the percentage identity. C) Alignment of the N-terminus of a simulated protein based on Dbp1/Ded1 using our ‘realistic’ simulation of evolution (see Methods). D) Histogram shows the p-value distribution obtained from set of protein sequences that were evolved as in C). Grey shaded area indicates the expected proportion of tests. Circles indicate the distribution of p-values obtained from the likelihood-ratio test described in A) when the test statistic is assumed to be chi-squared distributed (black circles) or non-central chi-squared distributed (white circles, “corrected”).