Structural Variant identification on pangenome graphs

Question

[Panos]

I have a pangenome graph built (with minigraph) from a bunch of human genome assemblies and I've been trying to find structural variants using this pangenome graph.

I used gfatools, which produces a bed file with the bubble sequence, but it doesn't classify the SVs into different types (e.g. insertions, deletions etc). And also it doesn't directly mention in which individual(s) each bubble is found; it does mention the segment IDs, from which (I guess) you can get this information but it's extra work!

So does anyone know of any tool that takes a genome graph as input and outputs a list of SVs (preferably in VCF format)?

Thanks!

PS: I'm aware of the tools which take genome assemblies as input and output SVs (e.g. PAV, Assemblytics, SyRI). However, I'm not really interested in them!

Answer 1

[colindaven]

There might be something here for you : https://github.com/colindaven/awesome-pangenomes

Otherwise, I know creating a pangenome with minigraph-cactus will give you lists of variants in VCF format. You can't get from minigraph rGFA to GFA format, but if you can get a GFA you can try tools like odgi pav.