Defining and cataloging variants in pangenome graphs.
29.6 million genetic variants were identified in a human pangenome reference graph, revealing extensive diversity beyond the GRCh38 reference.
- Why it matters: Understanding structural variation and reference bias is crucial for accurate genomic analysis, especially in complex regions like HLA-A and RHD, which are difficult to analyze with linear references.
- What they did: The study defined pangenome variants using a reference tree within a pangenome graph, analyzing a draft human pangenome constructed with Minigraph and Cactus, and identified millions of variants, including 3.5 million absent from GRCh38.
- The result: This work enables more comprehensive variant detection, especially in tangled, multiallelic regions, and provides open-source tools and a catalog to improve genomic analyses across diverse human populations.