Fig. 3

Atypical evolution around nucleocyclic TSSs in testes. a. Average mononucleotide incidence changes from the common ancestor of Hd-rR and HNI to the current Hd-rR genome upstream and downstream of nucleocyclic TSSs (positive values: net increase, negative values: net decrease). The A/T incidence increases and C/G incidence decreases are quite small (−0.05 % ~ 0.05 %). The mononucleotide compositions are almost conserved from the ancestor to Hd-rR genome. The mononucleotide incidence changes are more pronounced in downstream regions than in upstream regions. b. The histogram shows “mutational flux”, average rates of base change from ancestral base X to current Hd-rR base Y upstream/downstream of nucleocyclic TSSs. c. For each mononucleotide Z (A, C, G, or T), the histogram shows the sum of mutational flux, average mutation rates X to Z (the flow into Z), where X is a mononucleotide other than Z, and the sum of Z to X (the flow out from Z) upstream (denoted “Up” in the figure)/downstream (Down) of nucleocyclic TSSs. d. Rates of average base change at positions around nucleocyclic TSSs (see all changes in Additional file 9: Figure S8a). A running average over 41-bp window is shown in Fig. 3d-e. The red arrow show the approximate positions of the first nucleosome dyads downstream of the nucleocyclic TSSs, indicating that C to T, and T to C rates have peaks at the first dyads and valleys at the first linkers (see also Additional file 7: Figure S10d for the liver case). e. We analyzed the apparent the nucleosome position-related mutational bias through analysis of mutational fluxes following alignment of first nucleosomes. After identifying the first nucleosome dyads downstream of TSSs according to the method described in Fig. 2e mutational fluxes are graphed in relation to the nucleosome dyad. First nucleosomes (red lines) and genomewide nucleosomes (black lines) show a significantly opposite trend around the dyads in the entire genome; namely, p = 0.02 % and 3.14 % for C to T and T to C, respectively by one-tailed Wilcoxon’s ranksum test