search

mhoban / rainbow_bridge

GNU General Public License v3.0
5 stars 2 forks source link

Combine separate (LULU- and non-) taxonomy collapse processes #51

Closed mhoban closed 8 months ago

mhoban commented 9 months ago

The lca process is currently called two separate times: once for your raw data and once if you want to use LULU output.

The thing is, the LULU output includes a map telling you how to collapse zOTUs, so you only really need to run it once and then filter the output by LULU's output map.

Maybe we should combine those into a single process that optionally includes the zOTU map from the output of the LULU process.

mhoban commented 9 months ago

I think the approach here is to collapse/combine the raw blast results using the LULU-curated zotus and then run the LCA (taxonomy collapse) script on that, rather than running LCA first and collapsing the resulting zotus based on the lulu output.

mhoban commented 8 months ago

This is basically addressed as of b324b39fc7f90d0c4724c8f3bb251cf281fa4957 (and some subsequent commits).

Now, the LCA process is run for all zOTUs and taxonomy-only output is given in output/taxonomy/lca*. The finalize process combines taxonomy with both curated and uncurated zOTU tables and puts it in output/final.