Genome-wide chromatin contact analysis can provide direct evidence of pairwise interactions but higher-order interactions such as multi-contact hubs and same cell (allele) interactions cannot be determined with short-read sequencing. Based on the third generation sequencer Oxford Nanopore Technology (ONT), there is no length restriction imparted during library preparation. First, we aimed to use this technology to sequence multi-contact concatemers that are generated after in situ proximity ligation assay. Importantly, the result of this experiment can ascertain the diversity of chromatin contacts. Hence, we proposed to systematically develop and use a method for single-molecule chromatin capture (SM-C) to map the cellular epigenetic and functional heterogeneity, and robustly benchmark the method compared to other short read methods. For future perspective, removing the requirement for PCR amplification, which allows for pseudo-single cell resolution in that we have direct evidence that this multiway chromatin contact must exist in at least one single-cell.
Our initial experiments from method development of SM-C on bulk iPSC cells qualitatively recaptured similar features to standard Hi-C data whilst at the same time elucidating multi-way contacts. Working with commercial partner, GeneBay we maximized the through of sequencing by adapting the library preparation from same number of input cells to the PromethION platform that gave much greater genome-wide coverage. Ultimately this provides more statistical power to support chromatin-contacts. Early on we describe that size-section using SPRI beads, an effective means to enrich the library for multi-way contact (i.e. multi-contacts > pairwise contacts). Quantitively the number of unique contacts is substantially below that of the publicly available 4DN datasets but this is more down to the depth of sequencing as the number of unique contacts per Gbp sequenced was substantially superior for SM-C. Further to this we have established a few QC metrics to understand the reproducibility, consistency to other established methods, etc. Using eigenvectors we have compared our own datasets to publicly available datasets to see some distance in the variation of the data for what should be for intense and purposes the same biological samples.
Further to this using native gDNA we were able to run the several tools to benchmark the specificity and sensitivity to detect CpG methylation (cytosine methylation to produce m5C) which will important for later objectives in this proposal to understand allele-specific chromatin contacts and how these are influenced by DNA modification. Methylation at such CpG islands can infer the expression of downstream genes. We used the guppy basecaller from Oxford Nanopore Technologies and run m5C calling afterwards using the state-of-art tool ‘remora’ also from ONT. Furthermore we initially stated we would compared this to bisulfite sequencing of our samples but to date this has not been achieved but will be a focus for the upcoming year. Furthermore we decided that with recent improvements in ONT sequencing accuracy it is not worthwhile to generate parallel datasets with PacBio Hi-Fi read platform give the dramatic decrease in throughout relative to PromethION.
Our Collaborative partners at McGill University have been working on the optimization of preparing cells for shipment to RIKEN for sequencing. This has involved transferring our standard operating protocol for generating fixed cells, permeabilizing steps, and handling to see the applicability to primary macrophages/monoctyes.