Probabilistic nucleosome scoring and protection/breakpoint peak maps of circulating cell-free DNA (CH01 sample pool)

This dataset provides genome-wide Probabilistic Nucleosome Scoring (PNS) signal tracks and associated peak calls derived from circulating cell-free DNA (cfDNA) sequencing data from the CH01 healthy donor pool, as described in “Probabilistic nucleosome scoring for high-resolution mapping of circulating cell-free DNA breakpoints and chromatin topologies” . PNS is a fragment-level, probabilistic framework that models nucleosome protection and cfDNA cleavage patterns by integrating fragment endpoints and lengths within the canonical nucleosome-associated size range (120–180 bp). Each fragment contributes a signed, mean-normalised signal reflecting dyad likelihood (positive signal) and cleavage structure (negative signal), producing a genome-wide track centred on a true zero baseline. This avoids the coverage-dependent distortions associated with sliding-window normalisation approaches and enables more accurate and complete identification of nucleosome positioning and cfDNA breakpoint features. The dataset includes: Coverage track (bigWig), normalised by dividing all values by the genome-wide mean coverage (81.76), and scaled ×100 PNS signal track (bigWig), normalised by dividing all values by the mean score of nucleosome protection peak calls (16.7644), and scaled ×100 Nucleosome protection peak calls (bigBed), scores (maximum PNS of region; column 5) derived from positive-signal regions, were normalised by dividing by the mean nucleosome protection peak score (16.7644), and scaled ×100. cfDNA breakpoint peak calls (bigBed), scores (minimum PNS of region; column 5) derived from negative-signal regions are reported as absolute values, and were normalised by dividing by the mean nucleosome protection peak score (16.7644), and scaled ×100. Following normalisation and scaling (×100), a value of 100 corresponds to the mean (i.e., 100% of the mean), with values above and below 100 representing proportional deviations relative to this reference. Nucleosome and breakpoint peaks are defined as contiguous regions of at least 50 base pairs above or below zero signal, respectively, allowing short interruptions (≤5 bp). These resources provide a high-resolution, baseline-stable representation of nucleosome organisation and cfDNA fragmentation across the human genome. They are intended to support analyses of nucleosome positioning, chromatin architecture, fragmentomic signal structure, and the design of cfDNA-based molecular assays, including applications sensitive to conserved breakpoint locations.Example command used to generate chromosome 22 outputs:python3 PNS_with_nucleosome_peak_calling.py -b Snyder_bams/CH01/*.bam --mode 167 --frag-lower 120 --frag-upper 180 -c 22 --max-duplicates 0 -o CH01_chr22_PNS --score-format wiggz --peak-format bed8BAM files included BH01, IH01 and IH02 (available from https://kircherlab.bihealth.org/download/cfDNA/), which are aligned to the hg19 (GRCh37) reference genome.