Population-scale Y chromosome assemblies reveal recurrent remodeling within constrained architectures

The human Y chromosome remains underrepresented in genomic resources because its repetitive architecture is difficult to resolve and poorly represented by any single reference. Here, we present 142 curated near-complete de novo Y chromosome assemblies spanning 98 terminal-marker-defined lineages across 17 haplogroup clades and 180,000 years of evolution, establishing a population-scale sequence and annotation resource for the chromosome. Across 100 quality-filtered AZFc intervals, recurrent inversions, deletions and complex rearrangements resolve into 12 structural haplotypes, while DAZ, RBMY and TSPY families show distinct paralogue- and array-specific evolution. Complete centromere and Yq12 sequences reveal extensive lineage-specific change within comparatively stable higher-order architectures, and haplotype-resolved methylomes show distinct epigenetic states among similar repeat copies. The reference panel reduces analytical bias: in testis Iso-Seq data, the best-matching assembly increased median DAZ transcript alignment identity from 91.7% to 99.9% and improved isoform annotation. Together, this resource enables analysis of Y-chromosome structural, evolutionary and reproductive variation.