With the increasing availability of high-quality genome assemblies, pangenome graphs emerged as a new paradigm in the genomics field for identifying, encoding,
and presenting genomic variation at both population and species levels. However, it remains challenging to truly dissect and interpret pangenome graphs via biologically informative visualization.
To facilitate better exploration and understanding of pangenome graphs towards novel biological insights, here we present VRPG, a web-based interactive visualization framework for pangenome graphs.
VRPG provides efficient and intuitive supports for exploring and annotating pangenome graphs along a linear-genome-based coordinate system (e.g., that of a primary linear reference genome) and
synchronized with other linear-assembly-based annotation features. Given the current surging adoption of pangenome graph in genomics studies, we believe VRPG will become a highly useful assistant
to help researchers to better explore pangenome graph as well as the biology behind it.
As demonstrations, we pre-shipped 4 reference-projected pangenome graphs (1 for yeast and the other 3 for human) with this web server. The yeast (Saccharomyces cerevisiae) pangenome graph was built upon 163 genome assemblies from 142 strains. The three human (Homo sapiens) pangenome graph were built by 90 genome assemblies from 46 samples.
The source code of VRPG is available at (https://github.com/codeatcg/vrpg) under the MIT open-source license, with which users can visualize their own reference pangenome graphs locally.
Zepu Miao, Jia-Xing Yue*. (2025) Interactive visualization and interpretation of pangenome graphs by linear-reference-based coordinate projection and annotation integration. Genome Research, (in press; doi: 10.1101/gr.279461.124; demonstration available at https://www.evomicslab.org/app/vrpg/; software available at https://github.com/codeatcg/VRPG) [LINK]
The yeast reference pangenome graph was constructed by Evomics Lab. This graph was built upon 163 yeast genome assemblies from 142 strains. Briefly, we took the S. cerevisiae reference genome (denoted as 'SGDref') retrieved from the Saccharomyces genome database (SGD) as well as 162 assemblies from our recently released S. cerevisiae Reference Assembly Panel (ScRAP; Ref: O’Donnell, et al. Nature Genetics, 2023) to construct reference-projected pangenome graph by using minigraph with the command ‘minigraph -cxggs -l 5000’. With the SGDref as the reference genome, we incrementally added those 162 assemblies of other S. cerevisiae strains into the graph according to their phylogenetic distances to SGDref. Regarding the haplotype tag, for the yeast reference pangenome graph, we used "HP0" to denote haplotypes of haploid or homozygous diploid strains, while using "collapsed", "HP1", and "HP2" to denote collapsed, or the two phased haplotypes of heterozygous diploid strains. The constructed pangenome graph has been deposited at Zenodo.
Human Pangenome Reference Consortium (HPRC) constructed three human pangenome graphs using Minigraph, Minigraph-Cactus, and PGGB respectively based on 90 genome assemblies from 46 samples. For the Minigraph graph, we used the version deposited by Dr. Heng Li at Zenodo. For the Minigraph-Cactus graph, we used the HPRC v1.1 version, accessible via this link. For the PGGB graph, we used the HPRC v1.0 version, downloadable from here. For these human reference pangenome graphs, the haplotype tag "0" was used to denote haplotypes of haploid samples or those where the haplotype is indeterminate due to mosaicism or collapsing, while "mat" and "pat" were used to denote the maternal and paternal haplotypes of the phased diploid samples.
To prepare users' own pangenome graph and its visualization with VRPG, please refer to the instructions on the Help page as well as the VRPG GitHub repository.