MEIRLOP

MEIRLOP performs motif enrichment analysis on ranked lists of peaks while controlling for sequence composition biases such as GC content, dinucleotide composition and CpG-associated biases to improve detection of transcription factor (TF) binding motifs.

Key Features:

• Logistic Regression Model: Employs a logistic regression model to assess motif enrichment by modeling the likelihood of motif presence as a function of sequence scores from ranked lists of peaks and covariates.
• Sequence Bias Control: Incorporates covariates for GC content, dinucleotide composition and CpG-associated biases to reduce confounding in motif enrichment analyses.
• Comparison with Existing Methods: Demonstrated equal or improved recovery of transcription factor binding motifs in differential regulatory regions, including interferon-beta–responsive sites.
• Broad Applicability: Applicable to diverse NGS assays and experimental designs, including differential acetylation ChIP-seq data and ENCODE TF ChIP-seq datasets.

Scientific Applications:

• Regulatory inference in stimulus response: Identify transcription factors driving transcriptional responses such as interferon-beta stimulation by detecting enriched motifs in differentially active regions.
• ChIP-seq peak sequence analysis: Pinpoint sequence features and candidate TF binding motifs near target proteins in differential acetylation and ENCODE TF ChIP-seq experiments.

Methodology:

Applies logistic regression on ranked lists of peaks with covariates for GC content and dinucleotide composition, using MOODS for motif scanning and statsmodels for statistical modeling.