B-063 Digital Microfluidic Platform for Single-Molecule DNA Methylation Profiling within Complex Specimens
Abstract
Aberrant DNA methylation is an early and pervasive hallmark of tumorigenesis and represents a promising biomarker class for noninvasive cancer detection. However, detection of rare, heterogeneously methylated tumor-derived DNA molecules in complex clinical specimens remains technically challenging. Conventional methylation-specific PCR is limited to predefined patterns, while next-generation sequencing lacks sufficient sensitivity and cost-effectiveness for routine screening of ultra-rare epialleles (<0.1%). We developed a digital microfluidic platform that enables highly parallelized, single-molecule methylation profiling using digital high-resolution melt (dHRM) analysis. The platform digitizes bisulfite-converted DNA into a microfluidic array comprising 40,000 nanoliter-scale wells, enabling absolute quantification of individual template molecules. Following digital PCR amplification with methylation-agnostic primers, parallelized high-resolution melt analysis is performed across the entire array. Sequence-dependent thermodynamic signatures allow discrimination of methylation density on a CpG-by-CpG basis without sequencing. A bioinformatic pipeline integrates locus-specific methylation density distributions across a multi-gene panel to generate a composite probability score. Analytical sensitivity, dynamic range, and clinical feasibility were evaluated using synthetic controls and DNA extracted from routinely collected Pap specimens. The platform demonstrated single-molecule sensitivity, detecting methylated epialleles at frequencies as low as 0.00005% (1 in 2,000,000 molecules) in a high-background setting. Digital partitioning enabled robust quantification across four orders of magnitude while preserving intermolecular methylation heterogeneity. In a proof-of-concept study of Pap specimens from ovarian cancer patients and cancer-free controls, methylation density profiling of a multi-gene panel achieved an area under the receiver operating characteristic curve (AUC) of 0.90. Notably, discriminatory performance was driven by quantitative assessment of heterogeneous epiallelic distributions rather than binary methylation calls. This digital microfluidic dHRM platform provides ultrasensitive, low-cost, and sequencing-free epiallelic profiling suitable for routine clinical laboratory implementation. By combining single-molecule resolution with thermodynamic sequence discrimination, the technology overcomes key limitations of existing PCR- and sequencing-based methylation assays. Ovarian cancer detection from Pap specimens demonstrates clinical feasibility, while the underlying platform is broadly adaptable to liquid biopsy applications and multi-analyte diagnostic panels requiring detection of ultra-rare molecular variants.