Hi-C data processing and A/B compartment calculation

Hi-C data processing was done by using Docker for 4DN Hi-C pipeline (v43, https://github.com/4dn-dcic/docker-4dn-hic). First, we aligned raw fastq files by BWA with ”mem -SP5M” to mouse genome (mm10). The aligned reads were sorted by run-pairsam-parse-sort.sh with ”COMPRESS_PROGRAM=lz4c". We removed duplicated reads and extracted UU/UR/RU pairs by pairtools with "dedup" and "select" options. We added juicer-style fragment information to pairs file by fragment_4dnpairs.pl. Finally, we filtered out read pairs with '(chrom1 == chrom2) and (frag1 == frag2)'  and read pairs with map quality < 10 by pairtools with "select". 

".hic" files were produced by juicer_tools.jar with "pre", then Hi-C eigen scores in each chromosome were calculated by juicer_tools.1.7.5 with  "-p eigenvector KR".  If the average gene density in regions with negative eigen scores were higher than those with positive eigen scores, we multiplied -1 to the eigen score.