Extract linear representation of paths in VG graph

Question

[Ian Fiddes]

I am trying to use VG effectively as a genome compression tool for collections of highly related genomes. A key component of this is that I need to be able to randomly access subregions of an input genome, or path. How can this be accomplished with VG?

So far, as a test case, I have built a graph out of a few sequences, converted it to a sorted XG format file, and have found that I can use vg find -p to select a subset of the graph, but how can I convert this back to a linear sequence?

Answer 1

[Jouni Sirén]

You can use vg paths -F to extract entire paths in FASTA format. By default, this extracts all paths in the graph. You can use option -p FILE to specify a list of path names in a file and option -Q PREFIX to extract all paths with the given name prefix.

In order to extract subpaths, you can combine vg paths with vg find. An example:

vg find -p 22:30000000-30100000 -x chr22.xg | vg paths -v - -F > output.fa

Option -v tells vg paths that the input is a vg file, and filename - means stdin.