Skip to main content
Fig. 1 | Genome Biology

Fig. 1

From: Comprehensive analyses of partially methylated domains and differentially methylated regions in esophageal cancer reveal both cell-type- and cancer-specific epigenetic regulation

Fig. 1

Identification of PMDs in esophageal samples by a sequence-aware multi-model PMD caller (MMSeekR). A A graphic model of the present study design. B Dot plots showing average methylation levels for all CpGs across the whole genome, CpGs within CGI promoters, common PMDs, SINE, LINE, and LTR in different samples. The annotations from Takai et al. [26] were used for CGI methylation quantification. C Development of a new PMD caller. The MethylSeekR α score measures the distribution of methylation levels in sliding windows with 201 consecutive CpGs across the genome. α score < 1 corresponds to a polarized distribution towards a high or low methylation level (that is, HMDs), while α score ≥ 1 corresponds to the distribution towards intermediate methylation levels (that is, PMDs). PCC shows the correlation between the predicted hypomethylation score based on a NN model, and the actual methylation level. A strong negative correlation indicates regions favoring PMDs, while weak/null correlation favors HMDs. D PCA analysis of 45 esophageal samples using the top 5000 most variable 30-kb tiles for the three PMD callers. E, F Representative windows showing PMDs successfully identified by MMSeekR but failed to be detected by either MethPipe (E) or MethylSeekR (F)

Back to article page

Genome Biology

ISSN: 1474-760X

Contact us