The methylation data were analyzed using the Infinium Human Methylation 450k BeadChip array following Illumina Methylation HD BeadChip protocol, as previously described by our group in the same set of patients [29 (link)]. The Beadchips were scanned with the Illumina IScanSQ system, and the image intensities were extracted with the Genome Studio 2.0. The methylation level was expressed as a beta (β) value—calculated as the intensity of the methylated channel divided by the total intensity β values range from 0 (unmethylated) to 1 (fully methylated). β values can be broadly interpreted as the percentage of CpG methylation.
Quality control and normalization of β values (BMIQ method) were performed using the Champ package for R. We filtered out probes with p-values above 0.01 and beadcount <3 in at least 5%, confounding starts (NoCG, SNPS, Multhit, and XY). For statistical analyses, β values were transformed in M values using the equation M value = log2 (β/1 − β), as described elsewhere [30 (link)]. Comparison of β values and M values methods for quantifying methylation levels by microarray analysis were performed. Further, we mapped the CpGs of interest located in genomic regions of the three targets validated by RT-qPCR using the integrative multi-omics database iMETHYL [12 (link)] (p. 2).
Quality control and normalization of β values (BMIQ method) were performed using the Champ package for R. We filtered out probes with p-values above 0.01 and beadcount <3 in at least 5%, confounding starts (NoCG, SNPS, Multhit, and XY). For statistical analyses, β values were transformed in M values using the equation M value = log2 (β/1 − β), as described elsewhere [30 (link)]. Comparison of β values and M values methods for quantifying methylation levels by microarray analysis were performed. Further, we mapped the CpGs of interest located in genomic regions of the three targets validated by RT-qPCR using the integrative multi-omics database iMETHYL [12 (link)] (p. 2).