Circular Chromatin Conformation Capture Sequencing (4C-Seq) and Analysis

4C-seq was adapted from previously described protocols with the following specifications.^{23, 24, 25} Ten million cell aliquots of lymphoblastoid cells were crosslinked by adding formaldehyde (MilliporeSigma) to a final concentration of 2% for 10 min before the reaction was quenched with glycine (Promega) at a final concentration of 125 mM. After lysis, chromatin was further released by douncing with a 21.5-gauge needle and digested using 1,500 U HindIII (NEB) per reaction at 37°C overnight. After enzyme inactivation, ligation, reverse crosslinking, DNA isolation, and purification, DNA was subjected to a second digestion by 50 U Csp6I (Thermo Fisher Scientific) per reaction at 37°C overnight. Enzyme inactivation, ligation, DNA isolation, and purification were next performed to create the 4C libraries. Each library was amplified by inverse PCR with primers containing multiplexed overhangs (SS_198F-SS_212R as reported in Table S2) using 50 ng of 4C template per reaction and purified by Qiaquick PCR purification kit (QIAGEN). Samples were submitted to The Biopolymers Facility in the Department of Genetics at Harvard Medical School (Boston, MA, USA) for quality control on an Agilent 2200 TapeStation D1000 HS ScreenTape and by SYBR qPCR assay. Samples were spiked with 30% PhiX prior to sequencing on an Illumina HiSeq 2500 sequencer using the HiSeq Single-Read Rapid Cluster Kit v2 (Illumina) for 100 cycles. At least 14 million pass-filter reads were acquired per sample.

4C-seq datasets were imported into the Galaxy-BioTeam Appliance and demultiplexed using Barcode Splitter (Galaxy v.1.0.0).²⁷ Reads were aligned to GRCh37/hg19 and a custom-made derivative chromosome 20 (der(20)) hg19 reference genome using Bowtie2 (Galaxy v.0.2). The der(20) was made using a Perl script that parsed through chromosomes 20 and 22 and concatenated parts of the chromosomes at the resolved breakpoints (Web Resources). Bam files were then analyzed using FourCSeq (v.1.16.0) with normal parameters established by the FourCSeq protocol with the exception of using the der(20) for some analyses.²⁸ The interactions between fragments were then mapped to a circular plot using Circos (v.0.69-6), by applying the relative coordinates of each fragment within the respective chromosome.²⁹ p-adjusted significance values were used to superimpose a colored heatmap onto the plot. These were then transformed into linear plots using Photoshop (version CC 2015) by converting the image from Rectangular mode to Polar mode.