genome graph generated by Minigraph

[yihangs]

The document says that the graph generated by Minigraph has no cyclic, which from my understanding, is a directed acyclic graph. However, when I checked the file hprc-v1.0-minigraph-grch38.gfa, I find that there are edges from one + node to a - node (e..g L s424632 + s757 -), so it does not look like a DAG to me. Can you help explain on that?

Thanks!

[glennhickey]

By having no "cyclic paths", the document means there are no labelled paths in the graph that are cyclic. The graph itself contains reversing edges and is therefore, like you point out, cyclic.

[yihangs]

Thanks! Do these graphs contain inter-chromosomal edges? Or they are actually a sum of distinct graphs, each representing one chromosome?

[glennhickey]

One component / chromosome (including unplaced contigs for grch38-based graphs)