Mots-Clés
pangenome graphs
structural variants
translocation variant
graph algorithms
sequence alignment
genome simulation
Description
Supervisors:
Raluca URICARU, MCf (Bordeaux), BKB team, LaBRI, Univ Bordeaux, https://www.labri.fr/bench-knowledge-and-beyond
& Claire LEMAITRE, DR Inria (Rennes), Genscale team, Centre Inria of Rennes University (IRISA), https://team.inria.fr/genscale/
Internship title: Characterization and detection of translocation variants in pangenome graphs
Keywords: pangenome graphs, structural variants, translocation variant, graph algorithms, sequence alignment, genome simulation
Address where the internship will take place: LaBRI, Campus de Talence, Université de Bordeaux
Internship description:
Context
In genomic research, Structural Variations (SVs), which include large-scale alterations such as insertions, deletions, inversions and translocations, play a significant role in genetic diversity and disease. They have been under-detected due to limitations in early sequencing technologies. At the same time, the acceleration of high-quality “telomere-to-telomere” genome assembly has shifted genomic analysis from relying on a single reference genome to more comprehensive pangenome models, such as pangenome graphs [1]. These graph-based models are particularly promising for the analysis of complex variants, such as SVs. Nevertheless, most published studies have predominantly focused on small polymorphisms or simple insertions and deletions that are easier to detect within these graphs [2].
Genomic variants among the genomes represented in the graph are typically identified by detecting topological motifs known as bubbles. A bubble can represent a wide range of variant types and their combinations. However, current bubble detection tools, such as VG deconstruct and BubbleGun [3], do not provide a detailed characterization of detected bubbles, particularly in terms of SV type and breakpoint position. At present, bubbles are reported with minimal information, usually limited to a reference genome position and the alternative walks in the graph corresponding to different alleles. While simple isolated SNPs, insertions, and deletions can be easily distinguished based on allele sizes, larger balanced variants (e.g., inversions or translocations) are more challenging to identify among the large number of unannotated bubbles.
Objectives
A first work conducted in the Genscale team characterized the specific bubble motifs generated by inversion variants in these graphs and proposed a tool to detect and annotate them in the output of bubble detection tools [4]. We propose here to conduct a similar study for another balanced SV type: translocations. A translocation is a rearrangement that moves a genomic segment (typically >1 kb) at another locus of the genome, in a cut-and-paste way.
This internship has the following objectives:
- to characterize the topological motifs generated by translocation events in pangenome graphs.
- to develop a method to identify and annotate such events in the graphs or in the output of bubble detection tools.
- to apply the method on real pangenome data.
The first step will include the simulation of genomes with various sets of translocation events (depending on the size of relocated segments, and distance of relocation), then the construction of pangenome graphs with several existing pangenome graph builders (minigraph, minigraph-cactus and PGGB), and finally the calling of bubbles.
The second step will consist either in parsing and analyses of the output of bubble callers, or the development of specific graph traversal algorithms.
The intern will program in Python or Rust and will be able to capitalize on the libraries and tools dedicated to pangenome graphs developed by the Genscale team (Gfagraphs, Pancat).
References:
- [1] Pangenome graphs and their applications in biodiversity genomics. Secomandi et al, Nat Genet 2025.
- [2] A draft human pangenome reference. Liao et al. Nature, 2023.
- [3] BubbleGun: enumerating bubbles and superbubbles in genome graphs. Dabbaghie et al, Bioinformatics, 2022.
- [4] Investigating the topological motifs of inversions in pangenome graphs. Romain S, et al. BioRxiv 2025.
- Pancat: https://github.com/dubssieg/pancat and https://github.com/dubssieg/rs-pancat-compare
- GfaGraphs: https://pypi.org/project/gfagraphs/