asl / BandageNG

a Bioinformatics Application for Navigating De novo Assembly Graphs Easily
GNU General Public License v3.0
112 stars 10 forks source link

[RFC] Load graphs and paths in 'vg' format #89

Open asl opened 2 years ago

asl commented 2 years ago

@ekg Do you think it will be useful?

ekg commented 2 years ago

I don't think so. It's equivalent to GFA. So is the odgi format and other custom indexes.

One thing that might help in reducing memory would be to load the GBZ format. Then you get compressed paths in low memory. That's also equivalent to GFA but then you're winning with the succinct data model.

On Thu, Jul 14, 2022, 13:47 Anton Korobeynikov @.***> wrote:

@ekg https://github.com/ekg Do you think it will be useful?

— Reply to this email directly, view it on GitHub https://github.com/asl/BandageNG/issues/89, or unsubscribe https://github.com/notifications/unsubscribe-auth/AABDQEN7MTCDWSH6YNYMSXDVT746XANCNFSM53R7EZBQ . You are receiving this because you were mentioned.Message ID: @.***>

asl commented 2 years ago

Well, it is equivalent, yes. However if things are already in 'vg' format then one would essentially skip the conversion from vg to GFA. Certainly it does not sense to support all plethora of possible custom formats here and there, however, I'm thinking that supporting few most widely used might improve the user experience.

We do support gzip-compressed GFAs now and after some recent improvements the memory consumption was significantly reduced, so I think we are ok so far. Unless you're having some examples where we might want to reduce memory consumption :)