Fig. 1

SEM pipeline with methylation predicts the effect of methylation on transcription factor binding affinity. A Using SEMpl all kmers with a PWM score under the TFM-P VALUE threshold are generated for the given transcription factor [18]. B SEMpl then generates all possible ‘in silico variants’ for each position of a transcription factor’s motif. These enumerated kmers are aligned to the genome in regions of open chromatin by DNase-seq, and the average ChIP-seq signal is determined for each alignment to generate SEMpl predictions for each base individually. SEMplMe then expands on this SEMpl output by adding WGBS to divide ChIP-seq signal peaks of C and G into the proportion of their signal affected by DNA methylation using a weighted sum. C SEMplMe output is displayed as all 6 nucleotides, including methylated C (M), and G opposite to methylated C (W), at every position along the motif. All values are displayed as log 2 and normalized to an endogenous binding baseline set to 0 (dark gray line). A scrambled baseline is also included (dashed gray line)