Long reads capture simultaneous enhancer-promoter methylation status for cell-type deconvolution.

Area of Science:

• Epigenetics and Genomics
• Molecular Biology
• Computational Biology

Background:

• Promoter methylation reinforces tissue identity, while enhancer methylation influences cell state and cancer.
• Long-read sequencing now allows analysis of methylation on enhancer-promoter pairs within single DNA molecules.

Purpose of the Study:

• To investigate if single-molecule enhancer-promoter methylation profiles can detect rare cancerous cells.
• To explore analysis methods for deconvolving cell-type mixtures using genome-wide enhancer-promoter methylation data.

Main Methods:

• Utilized long-read optical methylome data from cell lines (GM12878 and myoblasts).
• Identified over 100,000 enhancer-promoter pairs present on multiple DNA molecules.
• Developed and applied a mixture deconvolution methodology for proportional cell composition estimation.

Main Results:

• Accurate estimation of cell compositions in synthetic mixtures using both promoter and pairwise enhancer-promoter methylation analysis.
• Demonstrated that pairwise methylation analysis can resolve subtle differences between cell populations of the same cell type.
• Successfully generalized mixture deconvolution from different cell types to distinct populations within the same cell type.

Conclusions:

• Single-molecule enhancer-promoter methylation analysis is a powerful tool for cell mixture deconvolution.
• This approach holds promise for detecting rare cell populations, including cancerous transformations.
• The developed methodology and code are publicly available for further research.