MPCI: A novel metric for quantifying DNA methylation patterns in NGS data

Area of Science:

• Epigenetics
• Genomics
• Biomarker Discovery

Background:

• Epigenetic alterations, especially DNA methylation changes, are linked to various diseases.
• Traditional methods analyzing single CpG sites miss crucial fragment-level methylation patterns.
• Methylation haplotype analysis offers improved discrimination but has limitations in quantification.

Purpose of the Study:

• To introduce a novel metric, the Methylation Pattern Consistency Index (MPCI), for quantifying DNA methylation patterns.
• To address limitations in existing metrics for regional methylation analysis, particularly in complex samples like liquid biopsies.
• To evaluate MPCI's performance against established metrics like MHL and dMHL.

Main Methods:

• Development of the Methylation Pattern Consistency Index (MPCI) to capture consistent methylation patterns across sequencing reads.
• Utilized whole-genome bisulfite sequencing data for analysis.
• Benchmarking MPCI against MHL and dMHL using cell type differentiation, multi-tissue classification, and in-silico cfDNA spike-in detection.

Main Results:

• MPCI demonstrated superior performance in distinguishing closely related cell types (CD4 vs. CD8) with an AUC of 0.915.
• Achieved high accuracy (0.92) in multi-tissue classification and detected in-silico cfDNA spike-ins at 1% abundance.
• In a clinical liver transplant cohort, MPCI significantly outperformed dMHL in discriminating pre- and post-transplant cfDNA profiles (Accuracy: 0.868 vs. 0.768).

Conclusions:

• MPCI is a robust metric for quantifying methylation patterns, offering enhanced discrimination capabilities.
• The findings support MPCI as a reliable tool for epigenetic biomarker selection and diagnostic applications, especially in liquid biopsies.
• MPCI is available as an R function to facilitate its use in research.