PSHIC reveals allele-specific chromatin organization by phasing-free structural modeling

Resolving the distinct 3D chromatin architectures of parental chromosomes is critical for understanding diploid genome function, but many haplotype-resolved Hi-C contact-map methods remain constrained by their reliance on phased single-nucleotide polymorphisms (SNPs). This dependency treats reconstruction as a simple tag-and-sort problem, leading to massive data loss, extreme fragility to phasing errors, and inapplicability in SNP-sparse contexts. We introduce PSHIC, a computational framework that reveals this hidden chromatin organization by employing phasing-free structural modeling, directly from aggregated Hi-C data. PSHIC models chromosomes as continuous manifolds and solves the inverse problem of finding two homolog-specific structures whose combined contacts best explain the observed data. Comprehensive simulations across varying sequencing depths, SNP densities, and phasing error rates demonstrate that PSHIC consistently outperforms established methods, achieving superior accuracy (SCC ≥ 0.987), loop detection (F1≥ 0.952), and structural fidelity (DER ≤ 0.04). Applications to real human and mouse Hi-C datasets show that PSHIC recovers known allele-associated features in X-chromosome inactivation, genomic imprinting, and regulatory networks. Results align with orthogonal experimental approaches including 3D DNA-FISH, 4C-seq, and ChIA-PET. PSHIC democratizes diploid chromatin analysis by enabling robust reconstruction in non-model organisms and other SNP-limited or poorly phased contexts. PSHIC reconstructs parent-specific 3D genome organization from standard Hi-C data without phased genetic variants. It outperforms existing methods and recovers validated features of X inactivation, imprinting and gene regulation.