Efficient dynamic variation graphs

Abstract

Motivation Pangenomics is a growing field within computational genomics. Many pangenomic analyses use bidirected sequence graphs as their core data model. However, implementing and correctly using this data model can be difficult, and the scale of pangenomic data sets can be challenging to work at. These challenges have impeded progress in this field. Results Here we present a stack of two C ++ libraries, libbdsg and libhandlegraph, which use a simple, field-proven interface, designed to expose elementary features of these graphs while preventing common graph manipulation mistakes. The libraries also provide a Python binding. Using a diverse collection of pangenome graphs, we demonstrate that these tools allow for efficient construction and manipulation of large genome graphs with dense variation. For instance, the speed and memory usage is up to an order of magnitude better than the prior graph implementation in the VG toolkit, which has now transitioned to using libbdsg’s implementations. Availability libhandlegraph and libbdsg are available under an MIT License from https://github.com/vgteam/libhandlegraph and https://github.com/vgteam/libbdsg. Supplementary information Supplementary data are available at Bioinformatics online.

A cattle graph genome incorporating global breed diversity

“…Following the generation of alignments with CACTUS, we used a custom pipeline to detect nodes that were not present in the Hereford genome, ARS-UCD1.2, considered as the reference genome. We first used a custom python script and the libbdsg 54 library to extract the nodes not present in any Hereford paths. These nodes have then been screened for Nmers, and then misassembled regions detected by FRC_Align 30 on the two de novo assemblies here presented were discarded.…”

“…We generated a linear expanded genome with the purpose of providing an easy to use, expanded version of the cattle reference genome that is also easy to implement in current best practice pipelines. We extracted all nodes not present in the linear Hereford genome, but that were found in the other 4 assemblies considered using libbdsg (v0.3) 54 . Nodes were then labelled based on the genome in which they were found (i.e.…”

ODGI: understanding pangenome graphs

“…3.Their embedded paths are locally similar to each other. These properties are used to build efficient dynamic variation graph data structures (Siren et al, 2020;Eizenga et al, 2020a). Sparsity (1) allows us to encode edges E using adjacency lists rather than matrices or hash tables.…”

Genotyping common, large structural variations in 5,202 genomes using pangenomes, the Giraffe mapper, and the vg toolkit

“…This is not always identical to the original sequence graph, as some nodes and edges may not be visited by any haplotype. In order to support the handle graph interface 43 , we need some additional structures:…”