## For the "data processing step" fields:
- Read pairs were mapped against the hg38 reference genome (excluding haplotype chromosomes) with Bowtie2, parameters "-X 2000 --fr --end-to-end --very-sensitive".
- Alignments were filtered to retain only properly-paired reads with MAPQ >= 5 (samtools view -f2 -q 5).
- Coverage tracks were generated by converting paired-end data to a single fragment (bedtools bamtobed with "-bedpe" option then extracting columns for chromosome and first/last coordinate), converting to bedGraph (bedtools genomecov with “-bg” option), scaling to 10 million fragments per sample, then converting to bigWig format via UCSC utility bedGraphToBigWig.
- Peak calls for analysis were defined as regions called as peaks in at least 5 samples, where called peaks in individual samples were made via HOMER v4.10.3 (makeTagDirectory with parameters "-keepAll -read1 -fragLength150" and findPeaks parameters "-region -size 500 -minDist 1000 -L 0 -tbp 0 -inputtbp 0", considering only samples with at least 5 million usable fragments), ignoring peaks overlapping blacklist regions.
- Fragment counts per peak were collected with bedtools coverage "-counts" using the same single-fragment BED files constructed for the coverage tracks.


## For the "genome/build assembly" field:
- hg38


## For the "processed data files format and content" field:
- bigWig files for mappped fragment coverage.
- Counts per peak files are tab-delimited text files reporting peak range and number of overlapping fragments.