Measures single cell DNA methylation. Melissa introduces a method able to simultaneously cluster cells using genome-wide methylation patterns while learning methylation profiles of regions of interest. This program leans on a Bayesian hierarchical model and can be used for CpG methylation states forecasting, without the requirement of additional annotation data. Besides, it can also determine subsets of cells based solely on epigenetic state.
Provides solution for clustering sparse single-cell DNA methylation data. Epiclomal is based on a hierarchical mixture model which pools information from observed data across all cells and neighboring CpGs to infer the cell-specific cluster assignments. This tool is composed of two variants: EpiclomalBasic that imposes less structure to the model by assuming the true hidden epigenotypes follow the same distribution across all the genomic functional regions considered; and EpiclomalRegion that allows their distribution to vary across regions.