Species doesn't seem to be working in review page

[ACharbonneau]

@ACharbonneau commented on Thu Oct 28 2021

I submitted a dataset where every subject is labeled as human, and it is displaying properly in the data browser:

but my review page says all my subjects are Not Specified

---

@mschor commented on Thu Oct 28 2021

I know this doesn't seem helpful, but it works when I hit the page. @karlcz, I don't have a good handle on how someone gets back species data from deriva. Do Amanda's posts look correct to you? When I hit the DCC Review page, it looks like this:

---

@karlcz commented on Thu Oct 28 2021

I think this is probably a bug in the ingest process, as I see blank species info in the "level1_stats" table that is driving these charts.

---

@karlcz commented on Thu Oct 28 2021

Digging deeper, I think this is working as designed and you just have absurdist test data. You've combined the taxonomic term for homo sapiens as a pathogen in a microbiome subject!

The definition of species we are using for the stats and charts only considers a narrow profile where the subject has granularity=single organism and has exactly one taxon specified, which is for clade=species, role=single organism.

---

@ACharbonneau commented on Thu Oct 28 2021

Amanda needs to fix her script.

actually I think this is just a quirk of your synthetic data! you have granularity set to "microbiome" and role set to "pathogen".
Our definition for the derived "subject species" concept is to only consider taxonomy as species when multiple conditions hold:
the subject has granularity "single organism" (violated here)
the subject-role-taxonomy association is for role "single organism"
the taxonomic term has clade "species" (violated here)
there is exactly 1 such term for the subject