Related Experiment Video
Updated: Mar 17, 2026
14:56
Sample Preparation to Bioinformatics Analysis of DNA Methylation: Association Strategy for Obesity and Related Trait Studies
Published on: May 6, 2022
5.3K
Comparison of pre-processing methodologies for Illumina 450k methylation array data in familial analyses.
Emma Cazaly1, Russell Thomson2, James R Marthick1
1Menzies Institute for Medical Research, University of Tasmania, Private Bag 23, Medical Sciences Building 2, Hobart, TAS Australia.
Clinical Epigenetics
|July 19, 2016
Summary
A new normalization strategy using stratified quantile normalization (QN) and ComBat effectively processes human methylome data from the Illumina 450k array, even without reference samples. This method preserves crucial biological information for accurate analysis in family studies.
Area of Science:
- Epigenetics and Genomics
- Bioinformatics and Computational Biology
Background:
- Human methylome studies often utilize Illumina Human Methylation 450k array (450k array) technology.
- Existing analysis pipelines are challenged by experimental designs lacking matched control or normal samples.
- This study addresses the need for robust normalization methods in methylome analysis, particularly in family-based inheritance studies.
Purpose of the Study:
- To evaluate the performance of eight normalization pre-processing methods for human methylome data.
- To identify the most suitable normalization strategy for experimental designs without reference samples.
- To ensure the preservation of biological variation during data processing.
Main Methods:
- Tested eight normalization methods on 50 samples from four families using 450k array BeadChips.
- Employed qualitative (density, MDS, cluster plots) and quantitative metrics (ANOVA, median absolute differences, standard error) for assessment.
- Evaluated the preservation of biological information by analyzing the association between a known methylation quantitative trait locus (mQTL) and methylation.
Main Results:
- Stratified quantile normalization (QN) combined with ComBat demonstrated superior performance across qualitative and quantitative assessments.
- This strategy significantly reduced batch effects (ANOVA p-value from <0.01 to 0.97) and minimized differences between replicated samples.
- Biological information, including the association between a known mQTL and methylation, was successfully preserved (p=1.05e-05).
Conclusions:
- A combined strategy of stratified QN and ComBat is highly appropriate for analyzing methylome data when no reference sample is available.
- This approach effectively preserves biological variation, crucial for studies investigating methylome inheritance in families.
- The findings highlight the importance of selecting appropriate data processing methodologies for advancing methylome research.
