Disease prediction by cell-free DNA methylation

Area of Science:

• Biochemistry
• Genomics
• Computational Biology

Background:

• Cell-free DNA (cfDNA) analysis is a rapidly advancing field for disease diagnosis.
• Current methods often rely on sequence variants, limiting applications to mutation-rich diseases like cancer.
• Epigenetic information, particularly DNA methylation, presents a promising avenue for broader diagnostic utility.

Purpose of the Study:

• To review and discuss statistical methodologies for cfDNA epigenetic profiling in disease diagnostics.
• To focus on marker selection and prediction model construction strategies for DNA methylation data.
• To compare different analytical approaches using simulations and real-world data.

Main Methods:

• Comprehensive literature review of statistical methods for cfDNA epigenetic analysis.
• Development and comparison of marker selection techniques.
• Evaluation of various prediction model construction strategies.
• Application of simulation studies and real-world dataset analysis.

Main Results:

• Identified key statistical challenges and advancements in analyzing cfDNA methylation for disease diagnosis.
• Demonstrated the performance differences between various marker selection and model building approaches.
• Provided data-driven recommendations for optimal analysis pipelines.

Conclusions:

• Epigenetic cfDNA analysis, especially DNA methylation, significantly expands diagnostic potential beyond mutation-based methods.
• Careful consideration of statistical methods for marker selection and model construction is crucial for accurate cfDNA-based diagnostics.
• This work offers guidance for researchers and clinicians utilizing cfDNA epigenetics for disease detection.