Widespread Compensatory Evolution Conserves DNA-Encoded Nucleosome Organization in Yeast
A) The TSS-distal trinucleotide spectrum is modified. Shown are trinucleotide frequencies at TSS-distal high and low occupancy sequences. Compared to the distribution at TSS-proximal sequences, the low occupancy sequences contain more A/T trinucleotides and less G/C trinucleotides. B) TSS-distal low occupancy sequences lose A/T slowly and gain them rapidly. Shown are ratios of substitution rates in low vs. high occupancy sequences (Y-axis) plotted against the substitution rates at high occupancy sequences (X-axis). Each point represents the rate of one of four types of substitution (color coded) in loci flanked by the 5′ and 3′ nucleotide depicted inside the data point. A/T losing substitutions (red, pink) are consistently slower in TSS-distal low occupancy loci, with very similar dynamics to those observed in TSS-proximal sequences (compare Fig 2). A/T gaining substitutions (blue, cyan) generally occur more rapidly in low occupancy loci than in high occupancy loci. A/T gains in G/C flanking contexts are somewhat conserved, though not to the extent observed in TSS-proximal low occupancy loci. C) Averaged substitution rates. Shown are the rates of A/T-gaining and A/T-losing substitutions at TSS distal (bars) and TSS proximal (gray ticks) high and low occupancy sequences, averaged over all flanking contexts. D) Evolution of G+C content at different occupancy levels. Shown are average G+C contents in the extant S. cerevisiae genome and in the inferred common ancestor of S. cerevisiae and S. paradoxus, depicted for 10 levels of nucleosome occupancy (Methods). An overall conservation of G+C content is observed. Conservation is disrupted for low occupancy sequences, suggesting that some of the low occupancy TSS-distal sequences in S. cerevisae have decreased their G+C content recently.