NCM 2021: Mapping protein-DNA interactions genome-wide with DiMeLo-seq

Annie explained that commonly used methods for mapping protein binding (e.g. ChIP-seq, CUT&RUN) use coverage as a proximity for binding: higher read coverage suggests that more protein is bound at that location. However, these methods do not fully capture all features of protein binding (e.g. with short reads, information on joint protein binding at neighbouring sites is lost), or they do not detect protein binding at a truly genome-wide scale (e.g. with short reads, interactions cannot be mapped in repetitive regions). They also rely on amplification, so DNA methylation information is lost; yet as protein binding is often influenced by the methylation status of a binding site, being able to identify binding in the context of endogenous methylation ‘would be very powerful’.
Annie’s team developed DiMeLo-seq (Directed Methylation with Long-read sequencing) to address these limitations. This method relies on direct detection of methyl-adenine bases, introduced exogenously in this technique to label where a protein has bound, with nanopore technology. This enabled the team to investigate protein binding in repetitive regions (such as centromeres); additionally, as endogenous CpG methylation information is retained, they could investigate CpG methylation status ‘for free’. With long reads, joint binding events could also be identified, on single molecules. ‘So we have all of this information contained within the molecules that we are sequencing with nanopore’.

Read the bioRxiv pre-print and the protocol on protocols.io.