Fig. 4: Changes in Chromatin Accessibility status following Prdm3/16 deletion.
a Heatmaps of ATAC-seq data made with the R package tornado plot showing 5067 regions with increased chromatin accessibility (left panel) or 4577 regions with decreased chromatin accessibility (right panel) in representative individuals, n = 6 controls and n = 5 Prdm3/16ShhCreΔ/Δ samples. Accessibility determined by differential accessibility analysis with R package DiffBind using a Wald Test with a |log2 fold change| cutoff of 0.58 and a p-value < 0.05. b Genomic distributions of each ATAC peakset, regions with increased or decreased accessibility, as annotated by HOMER annotatePeaks.pl. c Motif enrichment with HOMER searching either regions with increased or decreased accessibility, showing the putative transcription factors binding within these regions, p-value < 0.05, determined by binomial test comparing frequency of a given motif under a set of peaks compared to a randomly generated background. d ROC (Receiving Operating Characteristic) Curve and PR (Precision-Recall) Curves generated using gkmSVM to compare sequences under the open ATAC peaks trained against the closed ATAC peaks in Prdm3/16ShhCreΔ/Δ epithelial DNA. An AUC near one indicates the model is able to distinguish differences in the sequences within the peak sets. Error bars show variability in model performance, depending on how the data set is divided into training and testing subsets. e The transcription factors associated with the highest 500 and lowest 500 weighted kmers identified under the peaks using gkmSVM were identified using Tomtom (FDR < 0.05) and were graphed. f The UCSC Genome Browser was used to visualize selected changes in promoter and enhancer chromatin accessibility observed in Prdm3/16ShhCreΔ/Δ epithelial cells in differentially expressed genes associated with AT2 cell maturation.
