New Schemes Ready - Githubissues

phac-nml / biohansel

Rapidly subtype microbial genomes using single-nucleotide variant (SNV) subtyping schemes

Apache License 2.0

25 stars 7 forks source link

New Schemes Ready #86

Closed DarianHole closed 5 years ago

DarianHole commented 5 years ago

Here are the TB and Typhi schemes

I've added an automatic metadata inclusion for the typhi scheme that can also take an additional metadata table. It is currently under main as I couldn't figure out how to get it to work under the const.py script. I can change that if needed.

Still to do: ~~- Create pytests for these new schemes~~ ~~- Put in the full tb scheme, it is currently a placeholder as I wait for the new one~~ ~- Test the tb scheme~

TB scheme works well and is accessed under -s "tb_speciation" as requested

Pytests may take me some time as I have not done them before but I'll get those up soonish hopefully.

dankein commented 5 years ago

Hi @peterk87 yes we'll need a metadata file to make the scheme hierarchical again (I still need to make this due to some recent changes, hopefully very shortly)

glabbe commented 5 years ago

@peterk87 We will want to add metadata to the SE scheme after we decide on a shorter nomenclature targeted at source attribution, in collaboration with the epis and the NML surveillance team. It's not ready yet, but it's coming...

DarianHole commented 5 years ago

Merging in the new development to get fix incompatibilities should any be found. Edit - There is already one with the tile -> kmer change.

DarianHole commented 5 years ago

With the following change, metadata can either be specified on the command line, or included with the code (as typhi currently is). This allows for the following 4 cases of metadata to work with the program:

No metadata included and none specified
Metadata included with code added and none specified
No metadata included and specified metadata added
Metadata included with the code is added and metadata specified is added and then they are combined together and combined with the results

I believe that this is the best way to allow metadata to be included along with allowing users to add more of their own. Let me know if anything else should be changed with it.

peterk87 commented 5 years ago

@DarianHole I think that's great that you're building in so much flexibility with the subtype metadata. It'll make biohansel more useful for a larger audience :+1: