search

vgteam / vg

tools for working with genome variation graphs
https://biostars.org/tag/vg/
Other
1.09k stars 193 forks source link

vg chunk output is .vg and not configurable #2715

Open ekg opened 4 years ago

ekg commented 4 years ago

Now that vg convert lets us go direct from GFA to handlegraph models, we should be able to work fully with those models.

Some tools default to producing .vg files. We need to pick a good default format, and make it configurable in these cases.

I would suggest using PackedGraph or HandleGraph format by default.

glennhickey commented 4 years ago

I think a lot of the tools making .vg files are really writing hash graphs now. vg chunk can further configure the output with -O. It uses this pattern: https://github.com/vgteam/vg/blob/master/src/subcommand/chunk_main.cpp#L603 https://github.com/vgteam/vg/blob/master/src/subcommand/chunk_main.cpp#L687

But peeking at the code, -O doesn't seem supported for the trace haplotype option which was maybe never fully handlified.

I'd like to see -O standardized across all tools that make new graphs, kind of like bcftools and samtools. No reason not to have GFA as an option for it, as well as to let handle graphs be loaded directly from GFA into some default internal format.