GetBlunted nearly atomizes Lancet GFAs

[adamnovak]

The Lancet team at NYGC wants to display their compressed de Bruijn graph GFAs on vgteam/sequenceTubeMap. However, when they run them through GetBlunted to bluntify them, the output GFAs seem to be made up of 1 and 2-base nodes, well in excess of the number that should really be required by the graph structure.

We need to figure out a way to bluntify their graphs while retaining a graph structure that is mostly made up of nodes of appreciable length.

It would also be good to be able to get and use a mapping back and forth between their input graphs and the bluntified output graphs, for (I think) converting read alignments and other annotations.

[adamnovak]

The second part might be handled by the provenance table, but the README doesn't explain the format that one should expect to see in that file.

[adamnovak]

It might make sense to try Stark here, since it is meant for dBGs specifically: https://github.com/hnikaein/stark

[rlorigro]

Thanks Adam, yes I could imagine GetBlunted potentially struggling depending on how the DBG looks. Let me know if that doesn't work for you, and if you are able to get an example dataset for this issue