Creation and Validation of the First Infinium DNA Methylation Array for the Human Imprintome [blood_imprintome]

Differentially methylated imprint control regions (ICRs) regulate the monoallelic expression of imprinted genes. Their epigenetic dysregulation by environmental exposures throughout life results in the formation of common chronic diseases. Unfortunately, existing Infinium methylation arrays lack the ability to profile these regions adequately. Whole genome bisulfite sequencing (WGBS) is the unique method able to profile these regions, but it is very expensive and it requires not only a high coverage but it is also computationally intensive to assess those To address this deficiency, we developed a custom methylation array containing 22,819 probes. Among them, 9,757 probes map to 1,088 out of the 1,488 candidate ICRs recently described. To assess the performance of the array, we created matched samples processed with the Human Imprintome array and WGBS, which is the current standard method for assessing the methylation of the Human Imprintome. We compared the methylation levels from the shared CpG sites and obtained a mean R2 = 0.569. We also created matched samples processed with the Human Imprintome array and the Infinium Methylation EPIC v2 array and obtained a mean R2 = 0.796. Furthermore, replication experiments demonstrated high reliability (ICC: 0.799-0.945). 17 whole blood samples analyzed with the Human Imprintome array with 2 replicates each

差异甲基化印记控制区（differentially methylated imprint control regions, ICRs）可调控印记基因的单等位基因表达。终生伴随的环境暴露所引发的其表观遗传失调，会导致常见慢性病的发生。遗憾的是，现有Infinium甲基化芯片（Infinium methylation arrays）无法对这些区域进行充分的甲基化图谱分析。全基因组亚硫酸氢盐测序（WGBS）是唯一可对这些区域进行甲基化图谱分析的方法，但该方法成本高昂，不仅需要高测序覆盖度，还需耗费较高的计算资源以完成相关评估。为弥补这一缺陷，我们开发了一款包含22819个探针的定制化甲基化芯片。其中9757个探针可定位至近期报道的1488个候选ICRs中的1088个区域。为评估该芯片的性能，我们制备了经人类印记组芯片（Human Imprintome array）与WGBS处理的匹配样本——WGBS是当前评估人类印记组甲基化水平的标准方法。我们对共享CpG位点的甲基化水平进行了比较，得到平均决定系数R²为0.569。我们还制备了经人类印记组芯片与Infinium甲基化EPIC v2芯片（Infinium Methylation EPIC v2 array）处理的匹配样本，得到平均决定系数R²为0.796。此外，重复实验证实该芯片具有较高的可靠性，组内相关系数（intraclass correlation coefficient, ICC）为0.799~0.945。我们采用人类印记组芯片对17份全血样本进行了检测，每份样本均设置2次重复。