Safe Harbor Canino — processed genome-wide tracks (ATAC-seq, RRBS, Hi-C)

Safe Harbor Canino — processed genome-wide tracks (ATAC-seq, RRBS, Hi-C) for ROS_Cfam_1.0, with the audit results of 12 October 2026. Companion dataset to the code archive (concept DOI 10.5281/zenodo.21996452; repository github.com/stardrako-create/SafeHarborCanine). Files ATAC_*: weighted-mean accessibility, signal variability and consensus peak frequency across 71 dogs (Jin et al. 2024, PRJNA1048909). ATAC_ehsan5_* covers 5 additional healthy-control dogs and ATAC_joined76_* combines the 71 and 5 dogs by cohort size (details below). RRBS_*: weighted-mean CpG methylation, coverage frequency and variability across the same 71 dogs (PRJNA1049514). HiC_*: multi-resolution contact matrix (.mcool), TAD insulation scores (100/250/500 kb windows) and raw insulation table, from Hi-C on a single individual (“Mischka”, Wang et al. 2021, PRJNA587469): intentionally low-weighted structural context, not a population-level track. METHODS.md: exact formulas, scripts and parameter values of the Mother Track construction. SafeHarborCanine-audit-results-20261012.zip: result tables, evidence and reports of the candidate audit (folders 07_FINAL_CANDIDATES_2026-09-19 and 08_CANDIDATE_CLOSURE_AUDIT of the repository). Start with 08_CANDIDATE_CLOSURE_AUDIT/FINAL_RESULTS_20261012.md. Audit status (12 October 2026) The defined computational review is complete for the data currently available. The shortlist is w01 (NC_051811.1:17255706–17256706; first priority for testing) plus one additional provisional candidate, ANO2-39708 (NC_051831.1:39708074–39709074). w11 (NC_051836.1:27041161–27042161) is not advanced because its structural context remains unresolved; this is not a demonstration that the locus is unsafe. bg1k_2848 and bg1k_0442 remain genomic comparators, not validated negative controls. Coordinates are ROS_Cfam_1.0, BED, zero-based and end-exclusive. These are computational candidates. Integration, expression and function have not been validated, and no guide RNAs are claimed. Earlier regional rankings and counts, including the 26 passing regions of earlier versions, are superseded. Scope of the tracks. The tracks come from the original per-cohort pipeline. Later versions of the pipeline (v1.20 onward) rebuilt the ATAC and RRBS tracks in a unified run; the code is in the repository, but the large rebuilt tracks are not archived in this record. Methodology — Mother Track construction The ATAC and RRBS tracks are not a naive average across the 71 dogs. Each uses a two-layer weighting scheme: A per-dog global QC weight (ATAC: FRiP + TSS enrichment; RRBS: mapping efficiency + bisulfite conversion rate), min-max normalized across the 71 dogs and rescaled onto [0.2, 1.0] so no dog is ever fully excluded. A per-bin local confidence, comparing each dog's own coverage/depth at that bin against that same dog's genome-wide background rate — this lets under-measured bins drop out of a dog's contribution instead of being averaged in as “closed” or “unmethylated.” Consensus ATAC peaks require support from a strict majority of dogs (>=36/71, following ArchR's addReproduciblePeakSet() convention), not a signal-magnitude threshold. TAD boundaries (Hi-C, single individual “Mischka”) come from cooltools insulation at 25 kb resolution, computed independently at 100 kb/250 kb/500 kb windows, Li-thresholded, boundary called if any window flags it. Full formulas, exact scripts, and parameter values: see METHODS.md (uploaded alongside the data files in this record, and in the repository at 04_tracks_processadas/ROS_Cfam_1.0/METHODS.md). Extended 76-dog ATAC cohort (V9-B robustness check) Ehsan Valiollahi provided 5 additional canine ATAC-seq samples (GEO GSE278027, PBMC, 2 breeds; healthy controls, see the correction below). They were processed identically to the original 71 dogs and combined two ways: ATAC_ehsan5_*: standalone Mother Track built from just these 5 dogs (same two-layer QC-weight/local-confidence formulas, QC weights anchored to the 71-dog reference range). ATAC_joined76_*: the two Mother Tracks (71-dog + 5-dog) combined by cohort size, joined(bin) = (71*mean_71(bin) + 5*mean_5(bin)) / 76, not a naive re-pool of 76 equally weighted individuals. Consensus ATAC peaks were recomputed from scratch with bedtools multiinter across all 76 dogs' individual peak calls (threshold rescaled to >=39/76, the same majority-vote rule). Rerunning the candidate-scoring logic of that time (V9-B) against the 76-dog inputs, with every other input held constant, reproduced the same 26/461 surviving candidates and the same top-10 order as the 71-dog run, with scores shifting by at most ~0.0006: a robustness check of the scoring, not a new result. Later versions superseded those counts (see the audit status above). Correction (24 August 2026). An earlier version of this record described the five dogs as tumor-arm samples, based on a misleading study-wide GEO “tissue” field. The source paper's sample-naming convention (N_ = normal) confirms that all five are healthy controls. Their QC weights were also re-anchored to the 71-dog reference range instead of being normalized within the 5 dogs (all five sit at percentile 45–86 of the 71-dog FRiP/TSS range); the files in this record reflect the corrected weights. Neither correction changed the ranking result. Contributions. Pipeline design, implementation and audit: M. Paiva Sequeira. Initial idea and initial SHIP candidate generation: E. Valiollahi, who also provided an independent regulatory-element set and five additional control ATAC-seq samples. Supervision: V. M. Barreto. Parts of the code, analysis and documentation were produced with AI coding assistants (Claude Code, Codex) working under the author's direction; the scripts, receipts and SHA-256 manifest (scripts/verify_manifest.py) are the verifiable record.