Identification of regions of accessible chromatin is important for identification of candidate transcription factor binding sites as gene expression regulators. Various open chromatin mapping methods such as ATAC-seq and DNaseI-seq can be used to identify regions of accessible chromatin, which in combination with characterization of histone modifications such as H3K4me1 and H3K27ac, can identify putative regulatory regions such as enhancers. However, analyzing samples using multiple next generation sequencing methods can be challenging when sample quantity is limiting or when sample storage conditions cause nuclear disruption that precludes subsequent accessibility mapping approaches.

Therefore, prediction of chromatin accessibility from genomic information that can be measured from very small input quantities, and that is not sensitive to nuclear and chromatin disruption, would offer a valuable approach for broad identification of accessible chromatin. Numerous DNA methylome studies have previously revealed that regions of chromatin accessibility are associated with highly localized decreases in DNA methylation level. Consequently, purified genomic DNA contains a sensitive marker for chromatin accessibility in the form of DNA methylation state, which can be comprehensively mapped from very low input samples, including at the single cell level.

Here we present a computational method to accurately predict regions of accessible chromatin based on whole genome bisulfite sequencing data, which can be used on data from multiple species. Our method takes into account areas of low methylation, but also incorporates information about differentially methylated regions, which are indicative of the regions of accessible chromatin. We confirm our findings with matching mouse ATAC-seq and histone modification datasets. This tool provides an effective approach to extract genome chromatin accessibility information from purified DNA, allowing mapping of these genomic features from samples that are challenging for conventional accessibility assays, such as human post-mortem brain samples.