Figure 5.

Nucleosome positioning and AT-rich regions correlate with single-stranded DNA. (A,B) Strategy for genome-wide mapping of single-stranded DNA (SS-seq). (A) When KMnO₄ enters purified nuclei, it modifies unpaired pyrimidine bases and prevents reannealing, creating stable single-stranded (SS) DNA. To map non-B DNA, purified KMnO₄-treated DNA is digested with the single-strand-specific S1 nuclease, and exposed DNA ends are ligated to biotinylated hairpin adapters. Following DNA sonication to generate 300-bp fragments, biotinylated DNA ends are enriched on streptavidin beads, and a second set of adapters is ligated. After digestion of adapter hairpins, the library is amplified and sequenced. The same procedure is applied to naked genomic DNA as a control. Details of the method are in Supplemental Figure S8A. (B) The distribution of the difference in counts from KMnO₄-treated nuclear DNA and genomic control DNA is calculated genome-wide. The SS-seq signal represents the residual signal obtained by subtracting SS reads from naked genomic DNA from the SS reads from embryo DNA. Sites of KMnO₄ reactivity are identified by peaks in the SS-seq signal. (C–J) Multiparamteric comparison of observed SS-seq peaks and predicted single-stranded regions. Average profiles of SS-seq in the 2-kb region around observed (C) and predicted (G) SS DNA sites were plotted at 50-bp resolution. Average profiles of SIST-predicted probability of strand separation in the 2-kb region surrounding observed (D) and predicted (H) SS DNA sites. Nucleosome profile (MNase-seq) in the 2-kb region around observed (E) and predicted (I) SS DNA sites. Black arrows indicated well-positioned nucleosomes. Average GC content in the 2-kb region around observed (F) and predicted (J) SS DNA sites is plotted at 50-bp resolution. (K) Venn diagram shows intersection between predicted and observed SS DNA sites. (L,M) Heatmaps of GC content in the 2-kb region around observed SS-seq peaks and predicted SS DNA sites. SS-seq peaks and predicted sites were sorted by GC content; the surrounding sequences were aligned on top of each other and then binned (20 bp × 267 bp for L and 20 bp × 215 bp for M). Each bin was colored according to its GC percentage. (N,O) Heatmaps of SS-seq signal in the 2-kb region surrounding observed SS-seq peaks and predicted SS DNA sites that were sorted as in L and M. (P,Q) Heatmaps of MNase-seq in the 2-kb region around observed SS-seq peaks and predicted SS DNA sites, which were sorted as in L and M.