This work focuses on comparing pangenome constructions methods for the downstream implications of how well (and truthfully) variation is converted into currently-useful formats like VCF.
This reposistory is currently a work-in-progress as code is centralised and cleaned up.
This project involves multiple steps. We first
- construct pangenomes from 12 bovine assemblies with
- minigraph
- pggb
- cactus
- decompose pangenomes into vcf
- compare vcf files for accuracy to assembly-called truth
- assess applications to new topics like VNTRs.
The preprint associated with this work can be found here.