Related Experiment Video
Updated: Aug 24, 2025
14:56
Sample Preparation to Bioinformatics Analysis of DNA Methylation: Association Strategy for Obesity and Related Trait Studies
Published on: May 6, 2022
4.6K
Significant variation in the performance of DNA methylation predictors across data preprocessing and normalization
Anil P S Ori1, Ake T Lu2, Steve Horvath2,3
1Center for Neurobehavioral Genetics, Semel Institute for Neuroscience and Human Behavior, University of California Los Angeles, 695 Charles E. Young Drive South, Los Angeles, CA, 90095-176, USA. anilori.contact@gmail.com.
Genome Biology
|October 25, 2022
Summary
Choosing the right data processing is crucial for consistent DNA methylation (DNAm) predictor performance. This study evaluated 101 strategies, finding 78% of predictors achieved excellent consistency with optimal methods.
Area of Science:
- Epigenetics and Computational Biology
- Biostatistics and Bioinformatics
- Genomic Medicine
Background:
- DNA methylation (DNAm)-based predictors show potential for clinical applications.
- However, the consistency of their performance across different analytical pipelines is not well-established.
- Systematic evaluation is needed to understand how preprocessing and normalization strategies impact predictor reliability.
Purpose of the Study:
- To systematically evaluate 101 DNAm data preprocessing and normalization strategies.
- To assess the impact of these strategies on the consistency of 41 DNAm-based predictors.
- To identify optimal analytical strategies for reproducible DNAm predictor performance.
Main Methods:
- Analysis of a large EPIC DNAm array dataset (Jackson Heart Study, N=2053) with 146 technical replicate pairs.
- Evaluation of 101 distinct data preprocessing and normalization strategies.
- Assessment of predictor consistency using average absolute agreement between replicate pairs.
Main Results:
- 32 out of 41 (78%) DNAm predictors demonstrated excellent consistency with appropriate data processing.
- Moderate correlation in performance across strategies (mean rho=0.40) indicates significant algorithm heterogeneity.
- Effective removal of technical variation impacts downstream phenotypic association analyses, including mortality risk.
Conclusions:
- DNAm-based algorithms are sensitive to technical variation.
- Optimal data processing strategies are essential for reproducible estimates and improved prediction accuracy.
- This study provides guidelines for best-performing analytical strategies to enhance DNAm predictor performance and standardization.
