table S2.doc from DNA methylation subpatterns at distinct regulatory regions in human early embryos

DNA methylation has been investigated for many years, but recent technologies have allowed for single-cell and single-base resolution DNA methylation datasets and more accurate assessment of DNA methylation dynamics at the key genomic regions that regulate gene expression in human early embryonic development. In this study, the region staring from upstream 20 kb to downstream 20 kb of RefSeq gene was selected and divided into 12 distinct regions (up20, up10, up5, up2, 5'UTR, exon, intron, 3'UTR, down2, down5, down10 and down20). The candidate promoter region (TSS ± 2 kb) was further divided into 20 consecutive subregions, which were termed bins. The DNA methylation dynamics of these regions were systematically analysed along with their effects on gene expression in human early embryos. The dynamic DNA methylation subpatterns at the distinct genomic regions with a focus on promoter regions were mapped. For the 12 distinct genomic regions, up2 and 5'UTR had the lowest DNA methylation levels and their methylation dynamics were different with other regions. The region 3'UTR had the highest DNA methylation levels, and the correlation analysis with gene expression proved that it was a feature of transcribed genes. For the 20 bins in promoter region, the CpG densities showed a normal distribution pattern, and the trend of the methylated CpG counts was inverse with the DNA methylation levels, especially for the bin 1 (downstream 200 bp of the TSS). Through the correlation analysis between DNA methylation and gene expression, current study finally revealed that, the region bin −4 to 6 (800 bp upstream to 1200 bp downstream of the TSS) was the best candidate for the promoter region in human early embryos, and bin 1 was the putative key regulator of gene activity. This study provided a global and high-resolution view of DNA methylation subpatterns at the distinct genomic regions in human early embryos.DNA甲基化（DNA methylation）的研究已有多年历史，但近年来的技术进步使得单细胞单碱基分辨率的DNA甲基化数据集得以获取，并能更精准地评估人类早期胚胎发育中调控基因表达的关键基因组区域的DNA甲基化动态变化。本研究选取了覆盖RefSeq基因上下游各20 kb的区域，并将其划分为12个独立的区域：up20、up10、up5、up2、5'非翻译区（5'UTR）、外显子（exon）、内含子（intron）、3'非翻译区（3'UTR）、down2、down5、down10及down20。转录起始位点（Transcription Start Site, TSS）上下游±2 kb的候选启动子区域被进一步划分为20个连续的子区域，称之为分箱（bin）。本研究系统分析了这些区域的DNA甲基化动态变化，及其对人类早期胚胎基因表达的调控作用，并绘制了不同基因组区域（重点关注启动子区域）的动态DNA甲基化子模式图谱。在12个不同的基因组区域中，up2与5'非翻译区的DNA甲基化水平最低，且其甲基化动态变化模式与其他区域存在显著差异；3'非翻译区的DNA甲基化水平最高，结合基因表达相关性分析证实，该特征是转录活跃基因的典型标志。针对启动子区域的20个分箱，CpG密度呈现正态分布模式，甲基化CpG位点的数量变化趋势与DNA甲基化水平呈负相关，尤其是分箱1（转录起始位点下游200 bp处）。通过DNA甲基化与基因表达的相关性分析，本研究最终确定：分箱-4至6（转录起始位点上游800 bp至下游1200 bp的区域）是人类早期胚胎中启动子区域的最佳候选区域，而分箱1则是基因活性的潜在关键调控区域。本研究为人类早期胚胎不同基因组区域的DNA甲基化子模式提供了高分辨率的全局视图。