Fig. 3: Joint conditional probability density function to model the various positions of CRMs relative to each other.
From: Inferring chromatin architecture at a single locus through probabilistic in situ DNA localization
A Schematic illustrating how the three CRM pairwise distances were plotted as x-y coordinates to generate 3D contours. Probability Density Function (PDFs) of individual distances were combined to give rise to a 3D contour plot in which E1 is set at (0,0), E2 is on the x axis at (E1-E2, 0), and the PPE is plotted in the positive y quadrant (xPPE,yPPE). Grayscale bar along the x axis shows the E2 positional likelihood. B-B” PDFs for pairwise E1-PPE (B), E2-PPE (B’), or E1-E2 (B”) distances (μm) of WT, generated by gaussian kernel density estimation. Each distance data point is modeled as a probability distribution centered at the observed value, resulting in solid line traces. The shaded region surrounding line traces represents the 2.5% to 97.5% confidence interval. Adjacent heatmaps show two-sided Kolmogorov-Smirnov (K-S) test p-values (Benjamini–Hochberg corrected for multiple comparisons) using a color scale: yellow/pink is significant (p < 0.05 indicated by bracket and *), whereas purple/black is not; self-comparisons (black, p = 1.0) run diagonally through each heat map square. Arrows indicate occurrence of a significant change. C–F 2D projections of 3D landscape for WT over time, colored from white, purple, to orange to indicate low, medium, and high probability. G, H Subtraction contours comparing genotypes, generated by subtracting the second genotype (blue) from the first (red). Red indicates higher occupancy probability for the first genotype, blue for the second. The most probable position (MPP) of the first genotype is shown as filled circles, while the second genotype is shown as hollowed squares with their respective colors of blue (E1), green (PPE), and red (E2). All E1-E2 possible distances are shown below with the top grayscale being the first genotype. Distance modes for each triangle indicated by solid (contour 1) or dashed (contour 2) lines. Global two-sided Simes P-value and interpretations are shown for each subtraction contour. See also Fig. S3 and Source Data File for number of embryos and doublets for each condition, individual p-value as well as other relevant statistics.
