An Allele-Centric Pan-Graph-Matrix Representation for Scalable Pangenome Analysis
Abstract
Population-scale pangenome analysis increasingly requires representations that unify single-nucleotide and structural variation while remaining scalable across large cohorts. Existing formats are typically sequence-centric, path-centric, or sample-centric, and often obscure population structure or fail to exploit carrier sparsity. We introduce the H1 pan-graph-matrix, an allele-centric representation that encodes exact haplotype membership using adaptive per-allele compression. By treating alleles as first-class objects and selecting optimal encodings based on carrier distribution, H1 achieves near-optimal storage across both common and rare variants. We further introduce H2, a path-centric dual representation derived from the same underlying allele-haplotype incidence information that restores explicit haplotype ordering while remaining exactly equivalent in information content. Using real human genome data, we show that this representation yields substantial compression gains, particularly for structural variants, while remaining equivalent in information content to pangenome graphs. H1 provides a unified, population-aware foundation for scalable pangenome analysis and downstream applications such as rare-variant interpretation and drug discovery.
Summary
The paper introduces a novel allele-centric approach called the H1 pan-graph-matrix for representing and analyzing pangenomes at population scale. Existing pangenome representations often focus on sequence, paths, or samples, which can obscure population structure or fail to effectively compress data, especially for rare variants and structural variations. The H1 matrix treats each allele (SNV or structural variant) as a first-class object and encodes its exact haplotype membership using adaptive per-allele compression, choosing between dense bitmap and sparse list representations based on carrier frequency. A path-centric dual representation, H2, is also introduced, derived from the same allele-haplotype incidence data, restoring explicit haplotype ordering. The authors demonstrate the effectiveness of H1 and H2 using real human genome data from the 1000 Genomes Project. The key finding is that H1 achieves significant compression gains, particularly for structural variants, while remaining information-equivalent to pangenome graphs. The adaptive encoding strategy, switching between dense and sparse representations based on a defined break-even threshold (k* ≈ H / log2(H)), allows for near-optimal storage across both common and rare variants. This approach provides a unified and scalable foundation for pangenome analysis, enabling downstream applications such as rare-variant interpretation and drug discovery. The paper highlights the complementary nature of graph-based and matrix-based pangenome representations, emphasizing that the choice of representation should be driven by the specific analytical task.
Key Insights
- •The H1 pan-graph-matrix offers a novel allele-centric representation, shifting away from traditional sequence-, path-, or sample-centric views of pangenomes.
- •The adaptive per-allele compression strategy in H1, choosing between dense bitmaps and sparse carrier lists, is crucial for achieving near-optimal storage, particularly for structural variants.
- •The paper provides a theoretical break-even threshold (k* ≈ H / log2(H)) for determining when to use dense versus sparse encodings, where H is the number of haplotypes and k is the number of carriers of the allele.
- •H1 achieves a 78% reduction in storage compared to a bitmap-only representation for structural variants in a 2 Mb region on chromosome 1 using 400 haplotypes.
- •The authors introduce H2, a path-centric dual representation derived from H1, which restores explicit haplotype ordering while maintaining information equivalence.
- •The paper highlights the duality between graph-based pangenome representations and the pan-graph-matrix, emphasizing that they are information-equivalent but optimized for different analytical tasks.
- •The paper includes visualizations illustrating the impact of different levels of backbone segmentation on the complexity of pangenome graphs derived from the H1 pan-graph-matrix.
Practical Implications
- •The H1 pan-graph-matrix can be used as a foundation for scalable pangenome analysis, enabling efficient storage and querying of large genomic datasets.
- •Researchers and engineers can use H1 and H2 to improve the performance of rare-variant interpretation, drug discovery, and population stratification analyses.
- •The adaptive encoding strategy of H1 can be implemented in existing pangenome tools to improve compression efficiency, especially for datasets with a high proportion of structural variants.
- •The paper opens up future research directions in the development of application-specific pipelines and benchmarks for H1 and H2.
- •The separation of population incidence (H1) and haplotype ordering (H2) supports privacy-aware data sharing scenarios where sequence data are restricted or attached as external annotations.