Figure 5: Systematic simulated motif disruption reveals motif contributions to both initiation rate and TSS positioning.
(A) Schematic of the motif ablation process. (B) Example loci where individual motif instances were repeatedly ablated and the effect on ProCapNet predictions was estimated for each ablation. (C) The distribution of measured PRO-cap reads across all identified motif instances in PRO-cap peaks (top) vs. the predicted effects of ablating those motifs (bottom). FC, fold-change in predicted reads. Black bars indicate medians. (D) ProCapNet profile task predictions before vs. after motif ablation, aggregated (median) across all instances of each motif. Dashed line indicates where y = 0.
