EUVIMEDEUVIMED entdecken
Crossref

Measuring pangenome diversity at scale

Luca Parmigiani

PUB – Publications at Bielefeld University (Bielefeld University) · 05.10.2026

DOI: 10.4119/unibi/3020699 ↗

Kurz gefasst

Fachartikel in der Zeitschrift PUB – Publications at Bielefeld University (Bielefeld University), erschienen am 05.10.2026. Verfasst von Luca Parmigiani.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Pangenomes are defined as collections of genomes from the same species or genus. They offer a direct way to capture the genetic variation present within a species. Since their introduction in 2005, pangenomes have had a significant impact across many scientific fields, including drug design and plant breeding. With the growing availability of genome sequences, comparing them provides valuable context and can reveal changes in species characteristics that are not apparent at smaller scales. In this work, we propose a unifying framework for comparing pangenomes using sets of items, where an item can represent any genomic content of interest. We provide a mathematical foundation for this approach and explore different choices for items, including _k_-mers, which are DNA substrings of length _k_, and elements of pangenome graphs, that represent conserved regions as nodes and connect them with edges when they appear consecutively in some genomes. We focus on three key measures of pangenome diversity: openness, core size, and Hill numbers. Our main goal was to develop efficient methods to compute these measures and compare their values across different pangenome datasets. As part of this effort, we developed two software tools: Pangrowth and Panacus. Pangrowth calculates these measures directly from genomic sequences, while Panacus, co-developed with Daniel Dörr, computes them from pangenome graphs. Using _k_-mers, we analyzed pangenome openness in bacterial genomes, a measure indicating how rapidly a pangenome expands in terms of unique items. We found that our results are consistent with those obtained from gene-based methods. We introduced an exact and efficient approach to estimate pangenome growth, which is the expected number of items for subsets of genomes, without relying on repeated sampling, significantly improving computational performance. We also addressed the estimation of the core genome, defined as the set of items found in nearly all genomes, by modeling it as a function of the number of genomes. We explored pangenome graphs represented as colored compacted de Bruijn graphs, focusing on node count as a measure of structural diversity. This involved two key challenges: variability in node counts due to differing numbers of genomes and the disproportionate influence of rare genomic sequences. To address this, we developed a method for interpolating and extrapolating node counts based on genome numbers. Additionally, we applied Hill’s numbers, a diversity index from ecology, to balance the influence of rare and common nodes. Lastly, we explored a potential method for compressing pangenomes, with the aim of enabling future comparisons based on their compressed representations. Building on previous concepts such as eulertigs and matchtigs, which represent a set of _k_-mers as a spectrum-preserving string set with minimum cumulative length, we extended these ideas to 1-degenerate strings, where each position can represent a subset of the alphabet. We proved that finding such a representation is NP-hard for fixed _k_ ≥ 3 and for fixed alphabet size |Σ| ≥ 4, via two reductions from the minimum vertex cover problem.

Zitieren

APA

Parmigiani, L. (2026). Measuring pangenome diversity at scale. PUB – Publications at Bielefeld University (Bielefeld University). https://doi.org/10.4119/unibi/3020699

Vancouver

Parmigiani L. Measuring pangenome diversity at scale. PUB – Publications at Bielefeld University (Bielefeld University). 2026. doi:10.4119/unibi/3020699

BibTeX
@article{Parmigiani2026,
  author = {Parmigiani, Luca},
  title = {Measuring pangenome diversity at scale},
  journal = {PUB – Publications at Bielefeld University (Bielefeld University)},
  year = {2026},
  doi = {10.4119/unibi/3020699}
}
Kostenlos starten →