Skip to main content
Fig. 1 | BMC Bioinformatics

Fig. 1

From: A comparison of reference-based algorithms for correcting cell-type heterogeneity in Epigenome-Wide Association Studies

Fig. 1

Epigenetic Dissection of Intra-Sample Heterogeneity: the EpiDISH algorithm. a Given a tissue of interest and with knowledge of the main underlying cell subtypes, EpiDISH constructs a reference DNA methylation (DNAm) database for these cell subtypes using (i) DNase Hypersensitive Sites (DHS) and DNAm data for these cell types from existing public databases and (ii) a supervised selection procedure which identifies differentially methylated CpGs (DMCs) among each pair of cell-types. The resulting reference DNAm database is defined only over DMCs that map to a DHS in at least one of the underlying cell-types. b Given a tissue sample of interest, EpiDISH next infers underlying cell type proportions/weights using robust partial correlations, with non-negativity and normalization constraints imposed a-posteriori. Having estimated cellular proportions for each sample, feature selection against a phenotype of interest is then performed using these proportions as covariates

Back to article page