RELIC: A novel dye-bias correction method for Illumina Methylation BeadChip

Zongli Xu, Sabine A.S. Langie, Patrick De Boever, Jack A. Taylor, Liang Niu*

*Corresponding author for this work

Research output: Contribution to journalArticleAcademicpeer-review

Abstract

Background: The Illumina Infinium HumanMethylation450 BeadChip and its successor, Infinium MethylationEPIC BeadChip, have been extensively utilized in epigenome-wide association studies. Both arrays use two fluorescent dyes (Cy3-green/Cy5-red) to measure methylation level at CpG sites. However, performance difference between dyes can result in biased estimates of methylation levels. Results: Here we describe a novel method, called REgression on Logarithm of Internal Control probes (RELIC) to correct for dye bias on whole array by utilizing the intensity values of paired internal control probes that monitor the two color channels. We evaluate the method in several datasets against other widely used dye-bias correction methods. Results on data quality improvement showed that RELIC correction statistically significantly outperforms alternative dye-bias correction methods. We incorporated the method into the R package ENmix, which is freely available from the Bioconductor website (https://www.bioconductor.org/packages/release/bioc/html/ENmix.html). Conclusions: RELIC is an efficient and robust method to correct for dye-bias in Illumina Methylation BeadChip data. It outperforms other alternative methods and conveniently implemented in R package ENmix to facilitate DNA methylation studies.

Original languageEnglish
Article number4
JournalBMC Genomics
Volume18
Issue number1
DOIs
Publication statusPublished - 3 Jan 2017
Externally publishedYes

Keywords

  • Data preprocessing
  • DNA methylation
  • Dye-bias correction
  • Illumina HumanMethylation450
  • Illumina MethylationEPIC
  • RELIC

Fingerprint

Dive into the research topics of 'RELIC: A novel dye-bias correction method for Illumina Methylation BeadChip'. Together they form a unique fingerprint.

Cite this