Long-read sequencing was performed on 945 Han Chinese individuals. Structural variants were called per sample and then merged across all samples using SURVIVOR (v1.0.6). The merged VCF was converted to bigBed format for display. Allele frequencies and per-sample support information were extracted from the INFO fields of the merged VCF. The study identified two notable variants: an ancestral deletion in GSDMD associated with bone density and kidney injury risk, and a modern human-specific variant in WWP2 influencing height, body composition, and facial features.
The raw VCF data was obtained from the -OMIX +OMIX repository (accession OED00945268) at the National Genomics Data Center (NGDC), China National Center for Bioinformation.
The source VCF also encodes phased per-sample genotypes: the sampleList field on the detail page is derived from the SURVIVOR SUPP_VEC bitmask and is an ordered list of the 1-based indices of the 945 samples carrying each SV. The full per-sample phased VCF can be browsed as a separate track in the SVs from 945 Han Chinese entry of the Phased Variants track collection.
Thanks to Gong et al. for making their structural variant calls publicly available.