Haplotype-resolved karyotype construction from Hi-C data using refLinker

Chromosomal aberrations are prevalent in cancer genomes, yet it remains challenging to resolve the long-range structure of rearranged chromosomes. A key problem is to determine the chromosomal origin of rearranged genomic segments, which requires chromosome-length haplotype information. Here we describe refLinker, a new computational method for whole-chromosome haplotype inference using external reference panels and Hi-C. We show that refLinker ensures consistent long-range phasing accuracy in both diploid human genomes and aneuploid cancers, including regions with loss-of-heterozygosity and high-level focal amplification. We further demonstrate the feasibility of complex genome reconstruction using haplotype-specific Hi-C contacts, revealing new karyotype features in two widely studied cancer cell lines. Together, these findings provide a new framework for the complete resolution of long-range chromosome structure in complex genomes and highlight the unique advantages of Hi-C data for reconstructing cancer genomes with chromosome-scale continuity. ### Competing Interest Statement C.-Z.Z. is a scientific adviser for Pillar BioSciences. D.P. is a member of the Volastra Therapeutics scientific advisory board.