motif-site-occupancy-comparison

motif-site-occupancy-comparison

Summary

Compare transcription factor occupancy at known motif binding sites across two or more ATAC-seq conditions by applying footprint scoring and differential binding detection. This skill detects which transcription factors show altered chromatin occupancy between conditions by leveraging the visible depletion of Tn5 insertions around protein-bound sites.

When to use

You have bias-corrected ATAC-seq footprint signals (BigWig files) from two or more distinct conditions (e.g., different developmental stages, treatment vs. control) and a catalog of known transcription factor binding site motifs (in JASPAR, TRANSFAC, or similar format), and you want to identify which TFs show statistically significant changes in occupancy/binding intensity between conditions.

When NOT to use

• Input ATAC-seq data has not been corrected for Tn5 insertion bias; run ATACorrect first.
• You lack motif annotations or a reference TF binding site catalog; differential occupancy analysis requires known binding site locations.
• Your experiment has only a single condition or replicates that cannot be grouped; comparison requires at least two distinct biological conditions.

Inputs

• Bias-corrected ATAC-seq BigWig files (.bw) for each condition
• Peak region annotation (.bed file, typically from MACS2 or similar)
• Transcription factor motif catalog (.txt, .jaspar, or .pwm format)
• Reference genome sequence (.fa or .2bit)

Outputs

• Differential TF occupancy results table (.tsv) with TF names, footprint scores per condition, fold-changes, and p-values
• Visualization of top differential TFs showing occupancy changes and representative footprint patterns (.pdf or .png)
• BINDetect binding site predictions (.bed) with occupancy estimates per condition

How to apply

First, compute genome-wide footprint enrichment scores for each condition using the bias-corrected ATAC-seq signal and a set of reference motifs. Then apply differential binding detection (BINDetect) which compares footprint patterns at annotated TFBS motif sites between conditions, accounting for both footprint score changes and underlying sequence context. BINDetect produces a ranked list of TFs with differential occupancy estimates and p-values. Filter results by statistical significance (typically p < 0.05) and effect size thresholds to identify high-confidence differential TF occupancy events. Validate findings by visual inspection of aggregated footprint patterns and representative genomic tracks around top-ranking differential TF sites.