450k-array-data-processing

450k-array-data-processing

Summary

Comprehensive quality control and preprocessing pipeline for Illumina HumanMethylation450 array data, encompassing probe filtering, normalization, batch correction, and type-2 probe adjustment to produce a curated methylation matrix suitable for downstream statistical analysis.

When to use

Apply this skill when you have raw .idat files or beta-valued matrices from HumanMethylation450 (450k) arrays and need to remove low-quality probes, correct for technical artifacts (batch effects, type-2 probe bias), and normalize the data before performing differential methylation analysis or DMR detection.

When NOT to use

• Input is already a normalized, batch-corrected, publication-ready feature table from a prior preprocessing study
• You are analyzing EPIC array data (use EPIC-specific parameters and annotation instead)
• Raw signal intensities or unnormalized .idat files are not available and cannot be recovered

Inputs

• .idat files from HumanMethylation450 array experiments
• Beta-valued methylation matrix (probes × samples)

Outputs

• Filtered and normalized beta-valued methylation matrix
• Quality control report (probe counts before/after filtering, bead count distributions)
• QC plots (detection p-value distributions, bead count histograms, batch effect visualizations)

How to apply

Load raw 450k array data using ChAMP's data import functions (from .idat files or a beta-valued matrix). Apply champ.filter() with default parameters to remove probes with detection p-value > 0.01 and probes with fewer than 3 beads in at least 5% of samples. Select a type-2 probe correction method (SWAN, Peak Based Correction, or BMIQ—BMIQ is default). Apply functional normalization or other available normalization approaches. If batch effects are detected via SVD analysis, apply ComBat correction. Generate QC plots and a filtered probe matrix documenting the number of probes retained and removed at each filtering step. The rationale is that these sequential steps remove unreliable measurements, correct systematic biases introduced by array chemistry and batching, and ensure the final methylation values are comparable across samples.