ChromaFactor: Deconvolution of single-molecule chromatin organization with non-negative matrix factorization

Area of Science:

• Genomics
• Molecular Biology
• Computational Biology

Background:

• Single-cell chromatin organization studies are crucial for linking genome structure to function.
• Analyzing single-molecule data is challenging due to inherent cellular heterogeneity.
• Understanding individual chromatin fiber contributions to bulk trends requires advanced methods.

Purpose of the Study:

• To introduce ChromaFactor, a novel computational approach for deconvolving single-molecule chromatin organization data.
• To identify salient primary components within complex datasets.
• To correlate these components with functional genomic phenotypes.

Main Methods:

• Developed ChromaFactor, a computational method utilizing non-negative matrix factorization.
• Applied ChromaFactor to single-molecule imaging datasets across various genomic scales.
• Analyzed the contribution of individual molecules to identified components.

Main Results:

• ChromaFactor effectively deconvolves single-molecule chromatin organization data.
• Identified primary components correlate significantly with active transcription, enhancer-promoter distance, and genomic compartment.
• Found that bulk trends in genome organization are driven by rare cell subpopulations, not uniform changes.

Conclusions:

• ChromaFactor is a robust tool for analyzing chromatin structure-function relationships at the single-molecule level.
• The approach highlights the role of specific rare subpopulations in driving functional genomic changes.
• Provides new insights into cellular heterogeneity and its impact on bulk genomic phenomena.