CHALM Methylation Quantification for Cell Heterogeneity

The patent segments the bulk cell population into individual cell representations by treating each sequencing read as originating from a single cell. Instead of calculating a single mean methylation level for all cells, the method divides the data into individual read-level measurements, each representing a distinct cell's methylation status. This segmentation enables the detection of heterogeneity among cells while maintaining computational feasibility through read-based analysis.

The patent applies local quality by differentiating between methylated and unmethylated CpG sites within genomic regions rather than treating all CpGs uniformly. The method calculates methylation levels specifically at promoter regions and other functional elements, assigning different weights and interpretations to CpGs based on their local genomic context and biological significance. This enables the preservation of biological function information while maintaining manageable measurement complexity.

The patent transitions from a population-averaged view to a single-cell resolution view by analyzing each sequencing read as representing an individual cell. This dimensional shift from bulk to single-cell level allows the method to preserve the relationship between methylation and gene expression that would otherwise be obscured by cellular heterogeneity. The approach maintains sufficient statistical power by leveraging the large number of individual reads while capturing cell-to-cell variation.