motif-enrichment-statistical-testing

motif-enrichment-statistical-testing

Summary

Statistical testing of transcription factor binding motif enrichment in differentially accessible chromatin peaks using position-weight matrices and background models. This skill quantifies whether specific DNA motifs are overrepresented in a peak set relative to a null distribution, producing enrichment scores and p-values to identify functionally relevant regulatory factors.

When to use

After identifying a set of differentially accessible peaks (via tl.diff_test or equivalent), when you need to infer which transcription factors may regulate the observed chromatin state changes. Apply this skill when you have peak coordinates, a motif database (such as CIS-BP), and seek statistical evidence of motif overrepresentation to guide downstream regulatory network analysis.

When NOT to use

• Peak set lacks sufficient genomic annotation or quality control filtering—validate peak reproducibility and signal-to-noise ratio before enrichment analysis.
• Motif database is not curated for the target organism or tissue; mismatched PWM libraries will introduce false positives and reduce interpretability.
• Input peaks are already annotated with regulatory elements (e.g., promoters, enhancers from ChIP-seq); use ChIP-seq peaks directly instead for higher resolution.

Inputs

• Peak coordinate set (GRanges or BED-like object)
• Differentially accessible peaks (output from tl.diff_test)
• Motif position-weight matrix (PWM) database
• Reference genome sequence (FASTA or indexed)

Outputs

• Motif enrichment table (rows=motifs, columns=motif ID, enrichment score, p-value, FDR)
• Motif occurrence map (motif coordinates within peaks)
• Statistical summary (global false discovery rate, significance threshold)

How to apply

Load differentially accessible peaks as a feature set into SnapATAC2. Retrieve motif definitions and position-weight matrices from a reference database (e.g., datasets.cis_bp for CIS-BP). Invoke tl.motif_enrichment on the peak set, which scans for motif occurrences within the differential regions and computes enrichment statistics by comparing observed motif counts against a background model (typically peaks from the genome or matched by length/GC content). The function returns a motif enrichment table with motif IDs, enrichment scores (e.g., log-odds or fold-change), and statistical p-values. Validate that all required columns are present and non-null, and interpret p-values as measures of the statistical significance of motif overrepresentation after multiple-hypothesis correction (e.g., Benjamini–Hochberg FDR).