NuRepress: Inferring Transcriptional Repressors from Phased Nucleosome Architecture

Area of Science:

• Genomics
• Molecular Biology
• Computational Biology

Background:

• Transcriptional repressor identification is challenging due to limitations in current methods focusing on accessible chromatin.
• Regulatory signals within repressive chromatin domains are often undercharacterized.

Purpose of the Study:

• To present NuRepress, a novel computational framework for predicting transcriptional repressors.
• To integrate repressive chromatin architecture, functional signatures, and transcriptional outcomes for repressor inference.

Main Methods:

• NuRepress identifies well-phased nucleosome arrays within repressive chromatin as structural units.
• The framework stratifies arrays into functional subtypes based on Tn5 cut signal patterns.
• Candidate repressors are ranked by synthesizing subtype efficacy, motif enrichment, and transcriptional dynamics.

Main Results:

• Well-phased nucleosome arrays show species-conserved organizational patterns linked to repressive efficacy.
• Transcription factors (TFs) exhibit distinct spatial preferences relative to nucleosome arrays.
• NuRepress identified stage-specific nucleosome organization changes linked to transcriptional remodeling in pancreatic cancer.

Conclusions:

• NuRepress offers a structure-aware strategy for repressor inference, moving beyond accessibility-centered approaches.
• The framework links nucleosome organization to transcriptional outcomes, enabling dissection of transcriptional repression.