Skip to main content
Fig. 1 | Genome Biology

Fig. 1

From: Toward a base-resolution panorama of the in vivo impact of cytosine methylation on transcription factor binding

Fig. 1

Overview of JAMS model. A At each genomic region i, the JAMS model considers the control tag count (left) or the pulldown tag count (right) as a combination of background and/or TF-binding signals at that position. B Each of these signals are then modeled as a function of accessibility (Ai), methylation (Mi), and sequence (Si) at each region i. C Schematic summary of the predictor features extracted for each genomic location and the outcome variables. D The specifications of the generalized linear model used by JAMS. E Comparison between the observed and predicted CTCF binding signal in HEK293 cells [17]. F DNA accessibility coefficients learned by the CTCF JAMS model; each dot corresponds to the effect of accessibility at a 200 bp-bin. G Sequence motif logos representing the TF-binding specificity learned by JAMS (left) and the effect of sequence on the background signal (right). JAMS motif logos are plotted using ggseqLogo [18], with letter heights representing model coefficients. H The known CTCF binding preference (based on SELEX [19]); SELEX motif logo was obtained from the CIS-BP database [20].

Back to article page