CHAS

Cell type identity is a major driver of epigenetic variation, making biological interpretation of bulk tissue epigenomes difficult. Here we present CHAS (cell type-specific histone acetylation score), an R package for the cell deconvolution of bulk brain H3K27ac profiles. CHAS is split into two independent algorithms. In the first, CHAS annotates peaks identified in bulk brain studies of H3K27ac to cell type-specific signals in four major brain cell types, and based on normalised signal intensities generates cell type-specific histone acetylation scores to act as a proxy for cell type proportion. In the second, CHAS implements non-negative matrix factorisation based on the RNA-seq cell deconvolution tool EPIC, to estimate cell type proportions within bulk H3K27ac samples.