Hello, I have a question about the nature of the output from vg filter using the nodes parameter to extract read mapping metadata. See: https://github.com/vgteam/vg/wiki/Getting-alignment-statistics-with-vg-filter

When used, node traversals of each read are provided, like so:

["178767782+", "178767783+", … "178767800+"]
["174358302+"]
["125003213-", "125003211-", … "125003201-"]
["169920629-"]

Some set up for the question:

Given a walk through a graph: 1+,2+,3+,4-,5+

The forward strand of the haplotype traverses these nodes in the order depicted in the walk: 1+,2+,3+,4-,5+

But the reverse complement strand of the haplotype traverses these nodes in reverse: 5-,4+,3-,2-,1-

They both reflect the same walk through the graph, just coming from different directions depending on which strand you are following. My understanding is that the W lines in GFA files reflect the forward strand traversal through the graph. Is this correct?

My main question:

Because a read can be sequenced from either the forward or reverse complement strand, is the output of nodes from vg filter depicting both forward and reverse complement traversals by the reads, or are the nodes data processed in some way to depict read traversal through the nodes from only the forward strand, regardless of sequencing origin?

Depending on the answer to the above question, this read ["125003213-", "125003211-", … "125003201-"] could be interpreted as having been sequenced from the forward strand on a haplotype that traverses these nodes as an inversion (hence -) OR it could mean the read was sequenced from a reverse compliment strand, and the read "saw" these nodes in the opposite direction compared to how it is recorded in the GFA for the haplotype it mapped to (which would be ["125003201+", ... "125003211+", "125003213+"])

Thanks for your help!