Targeted high-resolution chromosome conformation capture at genome-wide scale

Abstract

24Chromosome conformation capture (3C) provides an adaptable tool for studying diverse 25 biological questions. Current 3C methods provide either low-resolution interaction profiles 26 across the entire genome, or high-resolution interaction profiles at up to several hundred loci. 27All 3C methods are affected to varying degrees by inefficiency, bias and noise. As such, 28 generation of reproducible high-resolution interaction profiles has not been achieved at scale. 29To overcome this barrier, we systematically tested and improved upon current methods. We 30show that isolation of 3C libraries from intact nuclei, as well as shortening and titration of 31 enrichment oligonucleotides used in high-resolution methods reduces noise and increases on-32 target sequencing. We combined these technical modifications into a new method Nuclear-33 Titrated (NuTi) Capture-C, which provides a >3-fold increase in informative sequencing 34 content over current Capture-C protocols. Using NuTi Capture-C we target 8,061 promoters 35 in triplicate, demonstrating that this method generates reproducible high-resolution genome-36 wide 3C interaction profiles at scale. 65The need to robustly sample these much more complex libraries has so far limited NG 66Capture-C to hundreds of viewpoints, generally performed in triplicate for statistical analysis. 67However, a large increase in the specificity of enrichment and the minimalization of off-target 68 and technical noise would practically translate into the feasibility of much larger viewpoint 69 designs using high-resolution methods. 71 4To dramatically increase the capacity of NG Capture-C, we have systematically optimized 72 multiple aspects of the protocol. 3C libraries in general are prone to technically induced noise, 73 which results in an increased frequency of non-informative trans reporters 9 . These spurious 74 reporters represent experimental background and so do not informatively add to the 75 interaction profiles, but do increase the required amount of sequencing. Consistent with 76 previous work 10 , we show that the 3C libraries can be separated into nuclear and non-nuclear 77 fractions with differing levels of information content. By optimising the 3C method to enrich for 78 and isolate intact nuclei after ligation we show a 30% increase in informative content. 80We next tested the effect of probe length and concentration on enrichment. Reducing 81 oligonucleotide probes from 120 to 50 bp resulted in a 5% increase in reads containing DpnII 82 sites. Additionally, titration of probe concentration resulted in a significant increase in capture 83 specificity, and when combined with double capture resulted in up to 98% on-target capture; 84 a 100-200% improvement over double capture alone. We have combined these optimisations, 85along with improvements to minimize losses during 3C DNA extraction and indexing 9 to 86 generate a modified protocol: Nuclear-Titrated (NuTi) Capture-C. 88The two seminal descriptions of targeted genome-wide 3C landscapes were carried out in 89 human ...

Multiplex-GAM: genome-wide identification of chromatin contacts yields insights not captured by Hi-C

“…Ligation is not the only potential source of difference between the two methods, as GAM and Hi-C also make use of quite different fixation protocols. The choice of fixation protocol has been shown to affect the proportion of informative ligation events between different 3C experiments (Oudelaar et al, 2017), and it may also influence the contacts of genomic regions with different protein composition and/or compaction within a single experiment (Downes et al, 2020;Williamson et al, 2014). The digestion of nuclear chromatin necessary for preparing Hi-C libraries has also been shown to disrupt nuclear structure (Gavrilov et al, 2013), whereas GAM uses fixation protocols specifically chosen to maximize the preservation of nuclear architecture and retention of nuclear proteins (Guillot et al, 2004).…”

Resolving the three-dimensional interactome of Human Accelerated Regions during human and chimpanzee neurodevelopment

“…As the DpnII fragment coordinates were different in human (GRCh38) and chimpanzee (panTro6), we generated different probe sets for each genome. Each DpnII fragment initially contained 2 probes at either end, and then filtered for repeat content (BLAT density score < 40), GC content (< 60%) and duplicate sequences (<=2 duplicates) using the program Capsequm 107,108 . Final probe sequences were generated on these filtered fragments using the design tool at https://capsequm.molbiol.ox.ac.uk/cgi-bin/CapSequm.cgi and pooled custom biotinylated oligonucleotide probes for each species were ordered from Twist Biosciences Inc.…”

The macro and micro of chromosome conformation capture

“…The NG Capture‐C method performs capture using DNA (rather than RNA) biotinylated oligos, introduces a second round of capture, and pools multiple independent 3C libraries for processing in a singular reaction, thus improving sensitivity and throughput. The latest iterations of Capture‐C design, Nuclear‐Titrated (NuTi) Capture‐C and Tiled‐C, further advance the resolution and efficacy of Capture‐C‐based methods (Downes et al, 2020 ; Oudelaar et al, 2020 ). NuTi Capture‐C isolates 3C libraries from intact nuclei by separating 3C libraries into nuclear and non‐nuclear fractions post‐ligation and utilizes shorter oligonucleotide probes (shortened from 120 to 50 bp).…”