Figure 3

p63 occupancy can be accurately predicted by chromatin marks. 5 active (H3K4me1, H3K4me2, H3K4me3, H3K9ac, H3K27ac) and 2 repressive (H3K9me3, H3K27me3) histone modifications with chromatin accessibility as measured by DNase-Seq and FAIRE-Seq were used to construct models of p63 binding. (A) Discriminant model for p63 binding. (B) Regression model for p63 occupancy. (C) The values for 11 datasets used for our models are shown with the data sorted in descending order by experimental p63 occupancy. The standardized number of tags for the histone modifications and chromatin accessibility are displayed. The DNase and FAIRE interaction term (DF) is calculated as product of the two and is displayed in standardized tag space. The quality of the p63 sequence motif was determined by PATSER, good motif matches are purple and low quality motif matches are pink. The results of our discriminant model is compared directly to the experimental determination of binding by MACS. Our regression model is compared directly to experimental occupancy.