Extended Data Fig. 2: Histogram of matched filter scores.
From: Supervised enhancer prediction with epigenetic pattern recognition and targeted validation
a) The probability density of matched filter scores for different epigenetic marks for STARR-seq peaks (positives) and random regions of the genome (negatives) with H3K27ac signal. In most cases, the matched filter scores for positives and negatives are Gaussian curves. The amount of overlap between these two curves determines the accuracy of the matched filter for predicting STARR-seq peaks using the matched filters for the corresponding epigenetic feature. b) The histogram of matched filter scores for small set of epigenetic features on promoters is compared to random regions of the genome. c) The histogram of matched filter scores for small set of epigenetic features on enhancers is compared to random regions of the genome. The features chosen in b, c were chosen to display distinct features of epigenetic marks around promoters and enhancers.
