![]() ( E) Composite profile of Rad21 binding at loop domains stratified based on the overlap with double, right, and no stripes as classified in activated B cells. Anchors were classified as not associated or associated with a stripe, either at the left, right, or both boundaries (double). ( D) Box plot shows Nipbl signals at loop anchors (+20Kb). ( C) Stripe (Hi-C) signals associated with anchors that display defective CTCF binding in ZF9-11 cells (red line) or no defect (black line). ( B) Example of loop domain and stripe loss when CTCF ZF9-11 fails to bind to loop anchors. Lower: representative ChIA-PET example showing the loss of loops in ZF9-11 cells at loci where CTCF binding is reduced. Related to Figure 3 ( A) Upper: APA analysis showing the cumulative Hi-C signals at loops where CTCF is either unchanged or >3 fold depleted in mutant cells. Stripe formation requires high-affinity CTCF binding. Right violin plots: Rad21 signals (ZF9-11/WT) at binding sites carrying various combinations of core (C), upstream (U) and downstream (D) motifs.ģ: Supplementary Figure 3. ChIP-Seq values are plotted as reads per kilobase per million (rpm). ( D) Representative example of defective binding of ZF9–11 mutant to upstream (U)-containing core (C) sites. ( C) Comparison of CTCF signals at left (blue line) and right (red line) anchors of loop (left graph) or stripe (right graph) domains, as determined by ChIP-Seq from activated B cells. An analogous approach was used to identify 5’ stripes, whereby the vertical to Left and vertical to Right ratios and p-values were considered (not shown). In the final step, loop contacts were removed from consideration (bottom right panel). Stripe calls were manually filtered using Juicebox visualization tool. ![]() Pixels with FC > 1.1 and p-value < 0.05 were retained and clustered to obtain stripes. In the horizontal stripe identification (bottom left panel, stripes from the left-hand side of the domains), the horizontal to Top and the horizontal to Bottom comparisons were considered. The 5’ and 3’ stripe identification (vertical and horizontal respectively) was separated at this stage. For each pixel, four values of fold changes and four associated P-values were obtained from the step above. Then stripes were identified and clustered as follows. Poisson statistics was used to assess whether the vertical or horizontal signal of the pixel was significantly higher than the local signal. For each pixel, four expected interaction strengths were obtained – given the Top, Bottom the Left and Right signal medians. Medians of the normalized signal in all the zones computed. Raw signal in the vertical and horizontal zone was aggregated. For each pixel separated by less than 3Mb (not shown), six local neighborhoods were defined which either included the pixel (vertical and horizontal zones) or were situated around it (Top, Bottom, Left and Right zones top right panel). Stripes appear as streaks of pixels displaying enriched signal as compared to the local background (top left panel, black box). ( B) Computational approach for stripe detection. Related to Figure 3 ( A) Examples of stripes (black arrowheads) in in situ Hi-C maps from human B lymphoblastoid cell lines (left), mouse ES cells (middle), or mouse CH12 B cells (right). Stripes are a new architectural feature that follows cohesin mobility in Hi-C maps. Data are represented as mean ± SEM.Ģ: Supplementary Figure 2. ( G) FACS analysis of Rad21-GFP recovery in HCT-116 cohesin degron cells where replication, transcription, or ATP synthesis was inhibited. ( F) Box plot comparing ChIP-Seq signals for WT Smc3, or mutants carrying E-Q or K-A substitutions. ( E) Cohesin “relative translocation” computed as the ratio of Rad21 or Smc3 ChIP-Seq signals at anchors/loading sites under the indicated conditions. Right ChIP-Seq tracks show a representative example of PolII profiles in the presence (upper) or absence (lower) of transcription. ( D) Cell cycle analysis (DAPI) of B cells activated for 30h in the presence of hydroxyurea (red histogram) or untreated (blue histogram). ( C) Histograms display the global distribution of Rad21 at Nipbl+ loading and CTCF+ anchor sites in G 0 (red) vs. ( B) Western blot showing expression of Rad21 (migrating at ~100KDa) and Actin control in activated B cells that are untreated or treated with oligomycin. Each point represents the integrated value of ChIP-Seq signals at Rad21 peaks. The datapoints are color coded based on CTCF signals. Related to Figures 1 and and2 2 ( A) Scatter plot shows a correlation between Nipbl and Rad21 signals genome wide. Cohesin distribution at loaders and anchor sites depends on ATP and cohesin ATPases.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |