High-quality structural variants have been discovered by long-read sequencing technologies across a few genomes. We use pangenome graphs to aggregate this information and genotype those variants accurately in short-read datasets, for example for population genetics studies or testing for their association with phenotypes.