Navigating the hydroxymethylome: experimental biases and quality control tools for the tandem OxBS Illumina microarrays. Supplementary data

<b>Table 1.</b> Multivariable linear regression model outputs for comparing DNA extraction kit performance between BS and oxBS treated samples, with sample source adjusted. <b>Table 2.</b> Linear mixed effect model outputs for comparing DNA extraction kit performance between BS and oxBS treated samples, with sample source included as a random intercept. <b>Table 3.</b> Pairwise Wilcoxon rank sum test p-values from comparing the internal normalization control probe (NORM_A/G/C/T) intensities in BS converted and oxBS converted ccRCC samples. <b>Figure 1.</b> The probe call rate difference across all data sets <b>Figure 2.</b> Principal component regression analysis of technical variables in ccRCC samples <b>Figure 3.</b> Coefficient of variation and ICC in replicate human fetal oligodendrocyte samples. <b>Figure 4.</b> The recommended 5hmC cut-off batch 1 samples batch 2 samples using <i>oxBSQC</i>. <b>Figure 5.</b> The distribution of 5hmC signal in low-quality probe vs high quality probe in batch 1 samples batch 2 samples using <i>oxBSQC</i>. <b>Figure 6.</b> The CpG genomic location enrichment for low-quality probes in batch1 and batch2 using <i>oxBSQC</i>. <b>Figure 7.</b> The distribution of internal normalization control probe (NORM_A/G/C/T) intensities in BS converted and oxBS converted ccRCC samples