MAGIC_Encode

MAGIC_Encode predicts transcription factors and cofactors that drive gene expression changes by detecting statistical enrichment in ENCODE ChIP-seq data (2314 tracks for 684 TFs and cofactors across 117 human cell lines) within gene bodies and their flanking regions.

Key Features:

• ENCODE ChIP-seq integration: Uses ENCODE ChIP-seq comprising 2314 tracks for 684 transcription factors and cofactors across 117 human cell lines to map factor occupancy within gene bodies and flanking regions.
• Enrichment-based prediction: Identifies statistical enrichment of TFs and cofactors within gene bodies and flanking regions rather than relying on binary target/non-target classification.
• Statistical approach: Implements a non-binary enrichment assessment that avoids reliance on Fisher Exact Tests used by traditional binary-target methods.
• Demonstrated contexts: Applied to cell lines with or without specific TFs, PAM50-classified breast tumors, whole-brain wild-type and TF-knockout mice, and single-cell RNA-seq neurons stratified by Immediate Early Gene expression.

Scientific Applications:

• Regulatory factor discovery in transcriptomics: Generates hypotheses about TFs and cofactors driving differential gene expression across experimental comparisons.
• Cancer subtype analysis: Links TF and cofactor occupancy to gene expression patterns in PAM50 breast tumor subtypes.
• Neurogenetics: Identifies neuronal transcription factors impacting whole-brain expression in wild-type versus knockout mouse models.
• Single-cell regulatory inference: Associates TF and cofactor enrichment with neuronal single-cell RNA-seq differences stratified by Immediate Early Gene expression.

Methodology:

Analyzes ENCODE ChIP-seq (2314 tracks for 684 TFs and cofactors across 117 human cell lines) to detect statistical enrichment of TFs and cofactors within gene bodies and flanking regions, assessing enrichment without binary target designation or reliance on Fisher Exact Tests.