Fig. 2
From: Estimating genome-wide DNA methylation heterogeneity with methylation patterns
Evaluation of methylation heterogeneity methods. A–C Estimating methylation heterogeneity with synthetic datasets. Top panel lists combinations of methylation patterns at different loci. Circles are model-based methods, and triangles are existing methods. Dashed lines represent the methods with increasing trends. Four types of scores are used in the comparisons, model-based methods: AB, PWS and PHY; accordance-based methods: MC, PDR, FDRP; Entropy-based: ME, and probability-based: EP; Existing methods considering pattern similarity: qFDRP and MHL. D The methylation heterogeneity of merged mouse ESC and muscle single-cell methylome estimating by PWS and ME. E Genome-wide methylation heterogeneity ratios are plotted against different numbers of ESC single-cell methylomes. The black line represents the expected values given merged cells are all heterogeneous while the red represents linearity