1. Methylome of human ovarian granulosa cells
Genomic DNA methylation patterns in ovarian granulosa cells were compared between NOR and POI. A total of 571.9 Gb of sequencing data averaging 95.3Gb per sample were obtained for the whole genome bisulfite sequencing of human ovarian granulosa cells. A statistical summary of various sequencing result parameters was reported. Clean reads sum up to average of 614 million read pairs. 0.3% of the cytosines were read as C instead of T. This indicated that the bisulphite conversion efficiency is 99.7%. The mapping efficiency for each library is 91.3%, 92.5%, 92.1%, 92.7%, 84.8% and 83.1% for C001, C002, C003, S001, S002 and S003 libraries, respectively. In our study the average of depth per strand was 13. The average of clean Q30 bases rate was 93%. An average of 582,490,035 cytosine methylation positions was called for each sample. DNA methylate cytosines can be classified into three types, CG, CHG, and CHH (H = A, G or T). The percentages of methylated cytosines in CG, CHG and CHH contexts were 99.15%, 0.19% and 0.65%, respectively for NOR and 99.24%, 0.15% and 0.58%, respectively for POI (Fig. 1). The average methylation level in CG, CHG and CHH were 80.14%, 0.4%, and 0.37% for NOR and 79.54%, 0.33%, and 0.31% for POI (Fig. 2). Genome-wide CpG methylation and CpG methylation level are generally similar between NOR and POI. We present Pearson correlation coefficient (PCC) to quantify the degree of colocalization between the six samples studied in this report (Fig. 3). There was a much greater similarity of DNA methylation pattern in NOR than in POI.
2.Human ovarian granulosa cells gene body methylation patterns
We analyzed the methylation levels of the promoters (2kb upstream), genes, Transcriptional Termination Region(TTR) (2kb downstream)(Fig. 4). The CpG methylation level approaching the TSS was a sharp drop. The CpG methylation increased rapidly after TSS and reached a plateau.There was no difference in trends but exited different methylation level in promoters between NOR and POI. To further investigate the CpG methylation patterns of human ovarian granulosa cells around the gene bodies, we extracted upstream, transcriptional start sites(TSS), first exon, first intron, internal exon, internal intron, last exon and downstream. A gene structure methylation distribution is shown in Fig. 5. Distribution of almost differentially methylated regions was in the context of upstream, TSS, first exon and internal exon.
3.Differentially methylated regions (DMRs) exited between NOR and POI
There was no significant difference in genome-wide CpG methylation between two groups. But there exited differentially methylated regions. We further analyzed distribution of DMRs across variou genomic elements such as the promoters, 5'-Untranslated Regions (5'-UTR), exon, Introns, 3'-Untranslated Regions (3'-UTR), CGI (Fig. 6). There was an obvious enrichment of methylation in exon. 5'-UTR had the minimum numbers. There exits correlation between methylation preference and sequence context in some plant [13]. In order to study a similar relationship in human granulosa cells, the distribution of methylated bases were calculated the standard deviation and mean of bases targeted by MTase motifs. Methylation percentage of 9-mer sequences in which the methylated cytosine was in the fifth position (allowing an analysis of three nucleotides upstream and four nucleotides downstream of CG, CHG, and CHH methylation) was calculated. The result showed that there was no obvious sequence context specificity in human granulosa cells, indicating that there is no correlation between sequence context and methylation preference (Fig. 7). So methylation preference couldn’t explain the formation of DMRs. In order to explore the biological function of differentially methylated genes related POI, we further performed the Gene Ontology (GO) Enrichment analysis. GO analysis was performed based on biological processes (BP), cellular components (CC), and molecular functions (MF). For BP, differentially methylated genes (sample VS control) tended to be enriched in the GO terms of cellular process, single organism process, biological regulation, metabolic process, developmental process. For CC, differentially methylated genes tended to be enriched in the GO terms of cell part, organelle part, membrane. For MF, differentially methylated genes tended to be enriched in the GO terms of binding and catalytic (Fig. 8).