"Gap hunting" to characterize clustered probe signals in Illumina methylation array data

Area of Science:

• Epigenetics
• Genomics
• Bioinformatics

Background:

• Illumina 450k arrays are crucial for epigenetic association studies.
• Existing quality-control (QC) pipelines remove probes with SNPs or poor mapping.
• Probes with bimodal DNA methylation distributions present additional QC challenges.

Purpose of the Study:

• To develop a data-driven method for identifying probes with clustered DNA methylation distributions.
• To investigate the relationship between probe characteristics and gap signals.
• To assess the utility of identified probes in downstream analyses.

Main Methods:

• Developed a "gap hunting" procedure to detect probes with clustered methylation patterns.
• Analyzed 590 peripheral blood samples from the Study to Explore Early Development.
• Paired genotype and 450k methylation data were used to validate findings.

Main Results:

• Identified 11,007 "gap probes" in the analyzed dataset.
• The majority of gap probes (9199) are associated with underlying SNPs or genetic variants.
• Factors like SNP type, probe type, and DNA strand influence gap signal occurrence.

Conclusions:

• Gap probes possess characteristics reflecting potentially informative biology.
• A data-driven approach to flag gap probes, rather than filter them, can enhance QC pipelines.
• Findings are applicable to the Illumina EPIC array due to similar design.