A validation of Illumina EPIC array system with bisulfite-based amplicon sequencing

PeerJ. 2021 Feb 10:9:e10762. doi: 10.7717/peerj.10762. eCollection 2021.

Abstract

The Illumina Infinium® MethylationEPIC BeadChip system (hereafter EPIC array) is considered to be the current gold standard detection method for assessing DNA methylation at the genome-wide level. EPIC arrays are often used for hypothesis generation or pilot studies, the natural conclusion to which is to validate methylation candidates and expand these in a larger cohort, in a targeted manner. As such, an accurate smaller-scale, targeted technique, that generates data at the individual CpG level that is equivalent to the EPIC array, is needed. Here, we tested an alternative DNA methylation detection technique, known as bisulfite-based amplicon sequencing (BSAS), to determine its ability to validate CpG sites detected in EPIC array studies. BSAS was able to detect differential DNA methylation at CpG sites to a degree which correlates highly with the EPIC array system at some loci. However, BSAS correlated less well with EPIC array data in some instances, and most notably, when the magnitude of change via EPIC array was greater than 5%. Therefore, our data suggests that BSAS can be used to validate EPIC array data, but each locus must be compared on an individual basis, before being taken forward into large scale screening. Further, BSAS does offer advantages compared to the probe-based EPIC array; BSAS amplifies a region of the genome (∼500 bp) around a CpG of interest, allowing analyses of other CpGs in the region that may not be present on the EPIC array, aiding discovery of novel CpG sites and differentially methylated regions of interest. We conclude that BSAS offers a valid investigative tool for specific regions of the genome that are currently not contained on the array system.

Keywords: Bisulfite based amplicon sequencing; DNA methylation; Illumina EPIC array.

Grants and funding

This study was funded by University of Otago Research Grant to Martin Kennedy, The Carney Centre for Pharmacogenomics. Christchurch Health and Development Study (CHDS) was funded by the Health Research Council of New Zealand (Programme Grant 16/600) and the Canterbury Medical Research Foundation. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.