ChAMP

ChAMP analyzes DNA methylation data from Illumina 450K BeadChip arrays to perform quality control, normalization, differential methylation analysis, and detection of differentially methylated regions for epigenome-wide association studies.

Key Features:

• Illumina 450K support: Processes methylation data generated by Illumina 450K BeadChip arrays.
• Quality control: Performs quality control procedures on methylation array data.
• Normalization: Implements normalization methods for 450K methylation data.
• Differential methylation analysis: Conducts differential methylation analysis between sample groups.
• Probe Lasso DMR detection: Implements the Probe Lasso algorithm that uses a flexible window-based strategy to aggregate neighboring significant probes and delineate DMR boundaries.
• Probe density adaptation: Adapts to varying probe densities to detect DMRs ranging from tens of bases to tens of kilobases.
• Probe filtering options: Allows probe filtering including inclusion or exclusion of sex chromosomes and polymorphisms.
• Adjustable Probe Lasso parameters: Permits modification of the probe-lasso size distribution according to study design.

Scientific Applications:

• Epigenome-wide association studies (EWAS): Enables EWAS by providing QC, normalization, differential methylation, and DMR detection workflows for 450K data.
• DMR discovery across scales: Detects DMRs from tens of bases to tens of kilobases, accommodating variable probe spacing.
• Cancer methylation analysis (TCGA): Has been applied to colon cancer and healthy colon samples from TCGA to shift DMR calling beyond densely populated probe regions.
• Transcription factor binding motif analysis: Facilitates identification of hypomethylated transcription factor binding motifs that may be missed by fixed-window approaches.

Methodology:

Computational methods explicitly include quality control, normalization, differential methylation analysis, probe filtering (sex chromosomes and polymorphisms), and the Probe Lasso algorithm employing a flexible window-based strategy that aggregates neighboring significant signals and adapts to varying probe densities; probe-lasso size distribution is adjustable.